{ "input": { "test_number": 99 }, "steps": { "input": { "test_number": 99 }, "build-batches": { "rwBatches": [ { "mod": 1, "type": "vocabulary_in_context", "count": 5, "startQ": 1, "testNumber": 99 }, { "mod": 1, "type": "text_comprehension", "count": 8, "startQ": 6, "testNumber": 99 }, { "mod": 1, "type": "data_analysis", "count": 4, "startQ": 14, "testNumber": 99 }, { "mod": 1, "type": "claim_support", "count": 4, "startQ": 18, "testNumber": 99 }, { "mod": 1, "type": "standard_english_conventions", "count": 5, "startQ": 22, "testNumber": 99 }, { "mod": 1, "type": "transitions", "count": 4, "startQ": 27, "testNumber": 99 }, { "mod": 1, "type": "rhetorical_synthesis", "count": 3, "startQ": 31, "testNumber": 99 }, { "mod": 2, "type": "vocabulary_in_context", "count": 5, "startQ": 1, "testNumber": 99 }, { "mod": 2, "type": "text_comprehension", "count": 8, "startQ": 6, "testNumber": 99 }, { "mod": 2, "type": "data_analysis", "count": 4, "startQ": 14, "testNumber": 99 }, { "mod": 2, "type": "claim_support", "count": 4, "startQ": 18, "testNumber": 99 }, { "mod": 2, "type": "standard_english_conventions", "count": 5, "startQ": 22, "testNumber": 99 }, { "mod": 2, "type": "transitions", "count": 4, "startQ": 27, "testNumber": 99 }, { "mod": 2, "type": "rhetorical_synthesis", "count": 3, "startQ": 31, "testNumber": 99 } ], "mathBatches": [ { "mod": 1, "type": "algebra", "count": 7, "spr": 2, "startQ": 1, "testNumber": 99 }, { "mod": 1, "type": "functions_analysis", "count": 5, "spr": 1, "startQ": 8, "testNumber": 99 }, { "mod": 1, "type": "geometry_trigonometry", "count": 5, "spr": 1, "startQ": 13, "testNumber": 99 }, { "mod": 1, "type": "statistics_probability", "count": 5, "spr": 1, "startQ": 18, "testNumber": 99 }, { "mod": 1, "type": "advanced_math", "count": 5, "spr": 2, "startQ": 23, "testNumber": 99 }, { "mod": 2, "type": "algebra", "count": 7, "spr": 2, "startQ": 1, "testNumber": 99 }, { "mod": 2, "type": "functions_analysis", "count": 5, "spr": 1, "startQ": 8, "testNumber": 99 }, { "mod": 2, "type": "geometry_trigonometry", "count": 5, "spr": 1, "startQ": 13, "testNumber": 99 }, { "mod": 2, "type": "statistics_probability", "count": 5, "spr": 1, "startQ": 18, "testNumber": 99 }, { "mod": 2, "type": "advanced_math", "count": 5, "spr": 2, "startQ": 23, "testNumber": 99 } ], "testNumber": 99 }, "gen-rw": { "count": 14, "costUsd": 1.91406, "failed": 0, "items": [ { "costUsd": 0.08664, "model": "opus", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 5 SAT Reading & Writing questions.\nType: vocabulary_in_context | Test: 99 | Unit: 1 | Start Q#: 1\n\nMANDATORY: Your JSON array MUST contain exactly 5 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nFollow the instructions for your type:\n\n[vocabulary_in_context]\n- 2-3 sentence passage with ______ blank from specific academic domains (molecular biology, Renaissance art, behavioral economics, astrophysics, political philosophy)\n- 4 college-level word choices from the same semantic field\n- Correct word is most PRECISE for context, not most common synonym\n- ID format: sat{testNumber}-rw-m{mod}-q{N}\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"vocabulary_in_context\",\"passage\":\"...with ______...\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[text_comprehension]\n- 3-8 sentence passage from published-quality source (journal, novel, essay) with specific names/dates/findings\n- Mix subtypes: main idea (2-3), author's purpose (1-2), inference (2-3), tone (0-1), structure (1-2)\n- Topics: natural science, social science, literature, history, arts/humanities (use each)\n- Distractors: surface-level misreading, wrong scope, opposite interpretation\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"text_comprehension\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[data_analysis]\n- Passage describing realistic data/study results requiring INTERPRETATION, not lookup\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Data patterns: diminishing returns, optimal values, conditional relationships\n- Distractors: wrong column, wrong trend direction, correlation vs causation confusion\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"data_analysis\",\"passage\":\"...\",\"visual\":{...},\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[claim_support]\n- Mix: \"Which finding supports/undermines the claim?\" AND \"Which quotation illustrates the claim?\"\n- Claims must be specific and testable. Distractors are topically related but don't address the specific mechanism.\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"claim_support\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[standard_english_conventions]\n- Passage with [underlined: text here] portion. Student picks best replacement.\n- Cover: subject-verb agreement, pronouns, verb tense, parallel structure, modifiers, commas, semicolons, apostrophes, sentence boundaries\n- \"NO CHANGE\" (choice A) correct ~25% of the time, NOT always\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"standard_english_conventions\",\"passage\":\"...with [underlined: ...]...\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[transitions]\n- Passage with ______ blank connecting two ideas. All 4 choices grammatically correct; distinction is logical.\n- Cover: contrast, addition, cause/effect, example, sequence, concession\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"transitions\",\"passage\":\"...______ ...\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[rhetorical_synthesis]\n- Notes format: \"While researching [topic], a student has taken these notes:\\n- Note 1\\n- Note 2\\n- Note 3\"\n- Specific rhetorical goal stated. All 4 choices are grammatical sentences using the notes; only 1 achieves the goal.\n- Goals: emphasize difference, highlight finding, introduce topic, summarize conclusion, present cause-effect\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"rhetorical_synthesis\",\"passage\":\"...\",\"prompt\":\"The student wants to [goal]. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\nOutput ONLY a valid JSON array of EXACTLY 5 objects. No markdown. Count them before outputting.", "text": "[{\"id\":\"sat99-rw-m1-q1\",\"number\":1,\"type\":\"vocabulary_in_context\",\"passage\":\"In his 1503 treatise on proportion, the Florentine sculptor argued that classical statuary achieved its effects not through rigid adherence to mathematical ratios but through subtle departures from them. A slight elongation of the torso, a marginal tilt of the head—these deviations, he claimed, ______ the figure with a vitality that strict symmetry could never produce.\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"imbued\",\"B\":\"contaminated\",\"C\":\"stocked\",\"D\":\"appointed\"},\"answer_format\":\"mcq\",\"correct_answer\":\"A\"},{\"id\":\"sat99-rw-m1-q2\",\"number\":2,\"type\":\"vocabulary_in_context\",\"passage\":\"Behavioral economist Sendhil Mullainathan has argued that scarcity does not merely shape decisions—it ______ cognition itself. When the poor must constantly calculate trade-offs that the wealthy never notice, mental bandwidth is consumed by the arithmetic of survival, leaving fewer resources for long-term planning.\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"designates\",\"B\":\"restructures\",\"C\":\"transcribes\",\"D\":\"deputizes\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m1-q3\",\"number\":3,\"type\":\"vocabulary_in_context\",\"passage\":\"The discovery of FOXP2, often misleadingly dubbed the ''language gene,'' has been widely ______ in popular science writing. Though mutations in the gene do disrupt speech production in humans, researchers caution that it is one component of a vast regulatory network rather than a singular switch governing linguistic capacity.\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"repudiated\",\"B\":\"curtailed\",\"C\":\"oversimplified\",\"D\":\"fabricated\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m1-q4\",\"number\":4,\"type\":\"vocabulary_in_context\",\"passage\":\"Because Population III stars—the hypothetical first generation formed from primordial hydrogen and helium—have never been directly observed, their properties must be ______ from indirect evidence such as the chemical signatures of ancient, metal-poor stars still visible in the Milky Way's halo.\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"extracted\",\"B\":\"inferred\",\"C\":\"fabricated\",\"D\":\"disclosed\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m1-q5\",\"number\":5,\"type\":\"vocabulary_in_context\",\"passage\":\"In $\\\\textit{On Liberty}$, John Stuart Mill worried that majority opinion, when unchecked, could exert a form of coercion more insidious than any statute. Public disapproval, he argued, could ______ dissent so thoroughly that heterodox ideas would wither before they could even be articulated.\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"stifle\",\"B\":\"rebuke\",\"C\":\"dispute\",\"D\":\"evade\"},\"answer_format\":\"mcq\",\"correct_answer\":\"A\"}]" }, { "costUsd": 0.280965, "model": "opus", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 8 SAT Reading & Writing questions.\nType: text_comprehension | Test: 99 | Unit: 1 | Start Q#: 6\n\nMANDATORY: Your JSON array MUST contain exactly 8 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nFollow the instructions for your type:\n\n[vocabulary_in_context]\n- 2-3 sentence passage with ______ blank from specific academic domains (molecular biology, Renaissance art, behavioral economics, astrophysics, political philosophy)\n- 4 college-level word choices from the same semantic field\n- Correct word is most PRECISE for context, not most common synonym\n- ID format: sat{testNumber}-rw-m{mod}-q{N}\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"vocabulary_in_context\",\"passage\":\"...with ______...\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[text_comprehension]\n- 3-8 sentence passage from published-quality source (journal, novel, essay) with specific names/dates/findings\n- Mix subtypes: main idea (2-3), author's purpose (1-2), inference (2-3), tone (0-1), structure (1-2)\n- Topics: natural science, social science, literature, history, arts/humanities (use each)\n- Distractors: surface-level misreading, wrong scope, opposite interpretation\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"text_comprehension\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[data_analysis]\n- Passage describing realistic data/study results requiring INTERPRETATION, not lookup\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Data patterns: diminishing returns, optimal values, conditional relationships\n- Distractors: wrong column, wrong trend direction, correlation vs causation confusion\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"data_analysis\",\"passage\":\"...\",\"visual\":{...},\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[claim_support]\n- Mix: \"Which finding supports/undermines the claim?\" AND \"Which quotation illustrates the claim?\"\n- Claims must be specific and testable. Distractors are topically related but don't address the specific mechanism.\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"claim_support\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[standard_english_conventions]\n- Passage with [underlined: text here] portion. Student picks best replacement.\n- Cover: subject-verb agreement, pronouns, verb tense, parallel structure, modifiers, commas, semicolons, apostrophes, sentence boundaries\n- \"NO CHANGE\" (choice A) correct ~25% of the time, NOT always\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"standard_english_conventions\",\"passage\":\"...with [underlined: ...]...\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[transitions]\n- Passage with ______ blank connecting two ideas. All 4 choices grammatically correct; distinction is logical.\n- Cover: contrast, addition, cause/effect, example, sequence, concession\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"transitions\",\"passage\":\"...______ ...\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[rhetorical_synthesis]\n- Notes format: \"While researching [topic], a student has taken these notes:\\n- Note 1\\n- Note 2\\n- Note 3\"\n- Specific rhetorical goal stated. All 4 choices are grammatical sentences using the notes; only 1 achieves the goal.\n- Goals: emphasize difference, highlight finding, introduce topic, summarize conclusion, present cause-effect\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"rhetorical_synthesis\",\"passage\":\"...\",\"prompt\":\"The student wants to [goal]. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\nOutput ONLY a valid JSON array of EXACTLY 8 objects. No markdown. Count them before outputting.", "text": "[{\"id\":\"sat99-rw-m1-q6\",\"number\":6,\"type\":\"text_comprehension\",\"passage\":\"In her 2019 study of cooperative breeding in pied babblers (Turdoides bicolor), zoologist Amanda Ridley documented that subordinate adults who had previously received vocal encouragement from dominant pair members were significantly more likely to provision unrelated nestlings than those who had not. Ridley argues that this pattern complicates the standard kin-selection account, which holds that helping behaviors evolve primarily because they benefit close genetic relatives. The vocal exchanges, she suggests, function less as instructions than as a kind of social currency that helpers accumulate and later redeem through opportunities to feed and be associated with the brood.\",\"prompt\":\"Which choice best states the main idea of the text?\",\"choices\":{\"A\":\"Ridley's research demonstrates that kin selection plays no measurable role in the cooperative behavior of pied babblers.\",\"B\":\"Ridley's findings suggest that social interactions, rather than genetic relatedness alone, may shape helping behavior in pied babblers.\",\"C\":\"Ridley discovered that dominant pied babblers use vocal commands to direct unrelated subordinates toward specific nestlings.\",\"D\":\"Ridley concludes that subordinate pied babblers provision nestlings primarily to gain reproductive opportunities of their own.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m1-q7\",\"number\":7,\"type\":\"text_comprehension\",\"passage\":\"The following text is adapted from Edith Wharton's 1905 novel The House of Mirth. Lily Bart, a young woman of limited means, has just encountered an acquaintance at a train station.\\n\\nSelden paused in surprise. In the afternoon rush of the Grand Central Station his eyes had been refreshed by the sight of Miss Lily Bart. It was a Monday in early September, and he was returning to his work from a hurried dip into the country; but what was Miss Bart doing in town at that season? If she had appeared to be catching a train, he might have inferred that he had come on her in the act of transition between one and another of the country-houses which disputed her presence after the close of the Newport season; but her desultory air perplexed him. She stood apart from the crowd, letting it drift by her to the platform or the street, and wearing an air of irresolution which might, as he surmised, be the mask of a very definite purpose.\",\"prompt\":\"Which choice best describes the function of the final sentence in the text as a whole?\",\"choices\":{\"A\":\"It resolves Selden's earlier confusion by clarifying Lily's reasons for being at the station.\",\"B\":\"It contrasts Lily's outward behavior with the more chaotic behavior of the surrounding crowd.\",\"C\":\"It introduces Selden's suspicion that Lily's apparent indecision may conceal a deliberate intention.\",\"D\":\"It suggests that Lily is unaware of Selden's presence and is preoccupied with private worries.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m1-q8\",\"number\":8,\"type\":\"text_comprehension\",\"passage\":\"Art historian Patricia Fortini Brown has long maintained that fifteenth-century Venetian painters approached classical antiquity differently from their Florentine contemporaries. Where Florentine artists such as Mantegna treated ancient ruins and inscriptions as authoritative models to be reconstructed with archaeological precision, Venetian painters like Gentile Bellini incorporated antique fragments into compositions that subordinated historical accuracy to civic narrative. For Brown, the resulting images are not failures of scholarship but deliberate acts of appropriation: Venice, lacking a Roman past of its own, fashioned one by absorbing the relics of others into a visual rhetoric of imperial continuity.\",\"prompt\":\"As used in the text, the phrase \\\"deliberate acts of appropriation\\\" most nearly suggests that Venetian painters\",\"choices\":{\"A\":\"borrowed antique elements carelessly because they lacked access to reliable archaeological sources.\",\"B\":\"intentionally repurposed classical material to construct a flattering historical identity for their city.\",\"C\":\"competed with Florentine artists by producing more historically accurate reconstructions of Roman monuments.\",\"D\":\"rejected classical sources entirely in favor of subjects drawn from Venetian civic life.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m1-q9\",\"number\":9,\"type\":\"text_comprehension\",\"passage\":\"Recent measurements from the Parker Solar Probe have revealed that the solar wind is punctuated by sharp, S-shaped reversals in the magnetic field, which mission scientists have termed \\\"switchbacks.\\\" Although these structures had been glimpsed by earlier spacecraft, their abundance close to the Sun was not anticipated by prevailing models of coronal heating. Justin Kasper, the mission's principal investigator, cautions that it remains unclear whether switchbacks are generated near the solar surface by small-scale magnetic reconnection events or arise farther out through turbulent interactions in the expanding wind. Distinguishing between these possibilities, he notes, will require coordinated observations from probes operating at different distances from the Sun.\",\"prompt\":\"Based on the text, what can most reasonably be inferred about the origin of switchbacks?\",\"choices\":{\"A\":\"Researchers have determined that switchbacks form primarily through magnetic reconnection at the Sun's surface.\",\"B\":\"Researchers consider switchbacks too rare near the Sun to be explained by current coronal heating models.\",\"C\":\"Researchers do not yet have observations sufficient to decide between competing explanations of how switchbacks form.\",\"D\":\"Researchers believe switchbacks are an artifact of the Parker Solar Probe's instruments rather than a genuine solar phenomenon.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m1-q10\",\"number\":10,\"type\":\"text_comprehension\",\"passage\":\"In The Federalist No. 10, James Madison concedes that factions, which he defines as groups united by interests adverse to the rights of others or to the permanent interests of the community, are an unavoidable feature of free societies. To suppress factions, he argues, would require either destroying the liberty that nourishes them or imposing a uniformity of opinion that is impossible among reasoning beings. Madison therefore turns from the question of preventing factions to that of controlling their effects, contending that an extended republic, by multiplying the variety of interests it contains, makes the formation of an oppressive majority less probable than it would be in a small, homogeneous democracy.\",\"prompt\":\"Which choice best describes the overall structure of the text?\",\"choices\":{\"A\":\"It defines a political problem, dismisses two proposed solutions, and advances an alternative remedy.\",\"B\":\"It presents a historical example, identifies its weaknesses, and proposes a return to earlier practices.\",\"C\":\"It introduces a controversial claim, surveys objections to it, and concludes that the claim cannot be defended.\",\"D\":\"It compares two competing forms of government and recommends a compromise between them.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"A\"},{\"id\":\"sat99-rw-m1-q11\",\"number\":11,\"type\":\"text_comprehension\",\"passage\":\"The following text is adapted from a 1903 essay by W. E. B. Du Bois.\\n\\nBetween me and the other world there is ever an unasked question: unasked by some through feelings of delicacy; by others through the difficulty of rightly framing it. All, nevertheless, flutter round it. They approach me in a half-hesitant sort of way, eye me curiously or compassionately, and then, instead of saying directly, How does it feel to be a problem? they say, I know an excellent colored man in my town; or, I fought at Mechanicsville; or, Do not these Southern outrages make your blood boil? At these I smile, or am interested, or reduce the boiling to a simmer, as the occasion may require. To the real question, How does it feel to be a problem? I answer seldom a word.\",\"prompt\":\"Which choice best describes the speaker's tone in the text?\",\"choices\":{\"A\":\"Wry restraint mingled with quiet weariness\",\"B\":\"Defiant outrage barely contained by politeness\",\"C\":\"Cheerful detachment about a trivial misunderstanding\",\"D\":\"Hopeful conviction that misunderstandings will soon be resolved\"},\"answer_format\":\"mcq\",\"correct_answer\":\"A\"},{\"id\":\"sat99-rw-m1-q12\",\"number\":12,\"type\":\"text_comprehension\",\"passage\":\"Behavioral economist Sendhil Mullainathan and his collaborators have argued that the experience of scarcity, whether of money, time, or calories, alters cognition in ways that conventional economic models fail to capture. In a series of experiments, participants asked to manage tight budgets performed better on tasks directly related to those budgets but worse on unrelated tests of fluid intelligence and self-control. Mullainathan describes this trade-off as a kind of cognitive \\\"tax\\\" levied by scarcity itself, one that may help explain why interventions designed to alleviate poverty often falter when they demand additional planning, paperwork, or attention from the very people they aim to assist.\",\"prompt\":\"Why does the author most likely describe scarcity as a cognitive \\\"tax\\\"?\",\"choices\":{\"A\":\"To suggest that governments should compensate individuals financially for the mental costs of being poor\",\"B\":\"To emphasize that scarcity imposes mental costs that can undermine programs requiring sustained effort from beneficiaries\",\"C\":\"To argue that conventional economic models accurately predict how people respond to financial constraints\",\"D\":\"To imply that people experiencing scarcity perform worse on every cognitive task than wealthier people do\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m1-q13\",\"number\":13,\"type\":\"text_comprehension\",\"passage\":\"Marine biologist Roger Payne's 1971 recordings of humpback whale songs are often credited with galvanizing public support for the international moratorium on commercial whaling adopted a decade later. Yet the recordings' influence cannot be attributed to scientific argument alone. By releasing the songs as a commercial album rather than confining them to academic journals, Payne placed the voices of whales within the same cultural space as popular music, inviting listeners to perceive the animals as expressive beings rather than as units of harvestable biomass. The conservation movement that followed drew as much from this aesthetic reframing as from the demographic data that researchers were simultaneously compiling.\",\"prompt\":\"Which choice best states the main purpose of the text?\",\"choices\":{\"A\":\"To document the scientific methods Payne used to record and analyze humpback whale vocalizations\",\"B\":\"To argue that the moratorium on commercial whaling was adopted primarily because of advances in population data\",\"C\":\"To suggest that Payne's recordings shaped public attitudes toward whales in part by reframing them aesthetically rather than only scientifically\",\"D\":\"To criticize the conservation movement for relying on emotional appeals rather than on rigorous scientific evidence\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"}]" }, { "costUsd": 0.15984, "model": "opus", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 4 SAT Reading & Writing questions.\nType: data_analysis | Test: 99 | Unit: 1 | Start Q#: 14\n\nMANDATORY: Your JSON array MUST contain exactly 4 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nFollow the instructions for your type:\n\n[vocabulary_in_context]\n- 2-3 sentence passage with ______ blank from specific academic domains (molecular biology, Renaissance art, behavioral economics, astrophysics, political philosophy)\n- 4 college-level word choices from the same semantic field\n- Correct word is most PRECISE for context, not most common synonym\n- ID format: sat{testNumber}-rw-m{mod}-q{N}\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"vocabulary_in_context\",\"passage\":\"...with ______...\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[text_comprehension]\n- 3-8 sentence passage from published-quality source (journal, novel, essay) with specific names/dates/findings\n- Mix subtypes: main idea (2-3), author's purpose (1-2), inference (2-3), tone (0-1), structure (1-2)\n- Topics: natural science, social science, literature, history, arts/humanities (use each)\n- Distractors: surface-level misreading, wrong scope, opposite interpretation\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"text_comprehension\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[data_analysis]\n- Passage describing realistic data/study results requiring INTERPRETATION, not lookup\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Data patterns: diminishing returns, optimal values, conditional relationships\n- Distractors: wrong column, wrong trend direction, correlation vs causation confusion\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"data_analysis\",\"passage\":\"...\",\"visual\":{...},\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[claim_support]\n- Mix: \"Which finding supports/undermines the claim?\" AND \"Which quotation illustrates the claim?\"\n- Claims must be specific and testable. Distractors are topically related but don't address the specific mechanism.\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"claim_support\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[standard_english_conventions]\n- Passage with [underlined: text here] portion. Student picks best replacement.\n- Cover: subject-verb agreement, pronouns, verb tense, parallel structure, modifiers, commas, semicolons, apostrophes, sentence boundaries\n- \"NO CHANGE\" (choice A) correct ~25% of the time, NOT always\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"standard_english_conventions\",\"passage\":\"...with [underlined: ...]...\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[transitions]\n- Passage with ______ blank connecting two ideas. All 4 choices grammatically correct; distinction is logical.\n- Cover: contrast, addition, cause/effect, example, sequence, concession\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"transitions\",\"passage\":\"...______ ...\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[rhetorical_synthesis]\n- Notes format: \"While researching [topic], a student has taken these notes:\\n- Note 1\\n- Note 2\\n- Note 3\"\n- Specific rhetorical goal stated. All 4 choices are grammatical sentences using the notes; only 1 achieves the goal.\n- Goals: emphasize difference, highlight finding, introduce topic, summarize conclusion, present cause-effect\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"rhetorical_synthesis\",\"passage\":\"...\",\"prompt\":\"The student wants to [goal]. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\nOutput ONLY a valid JSON array of EXACTLY 4 objects. No markdown. Count them before outputting.", "text": "[{\"id\":\"sat99-rw-m1-q14\",\"number\":14,\"type\":\"data_analysis\",\"passage\":\"Soil ecologist Maren Holst investigated how earthworm density affects the decomposition rate of leaf litter in temperate deciduous forests. She established experimental plots with varying numbers of earthworms per square meter and measured the percentage of leaf mass lost over a 90-day period. Holst hypothesized that decomposition would increase with earthworm density but that the rate of increase would diminish at higher densities because of resource competition among the worms.\",\"visual\":{\"type\":\"table\",\"title\":\"Leaf Litter Decomposition by Earthworm Density (90 Days)\",\"headers\":[\"Earthworms per m²\",\"Leaf Mass Lost (%)\",\"Increase from Previous Level (%)\"],\"rows\":[[\"0\",\"12\",\"—\"],[\"20\",\"28\",\"16\"],[\"40\",\"41\",\"13\"],[\"60\",\"49\",\"8\"],[\"80\",\"53\",\"4\"]]},\"prompt\":\"Which choice best describes data from the table that support Holst's hypothesis?\",\"choices\":{\"A\":\"Decomposition was highest at 80 earthworms per m², and the increase from the previous density level was the smallest observed.\",\"B\":\"Decomposition increased steadily from 0 to 80 earthworms per m², and the increase from each previous level was identical across densities.\",\"C\":\"Decomposition was lowest at 0 earthworms per m², and the increase from 0 to 20 earthworms per m² was smaller than the increase from 60 to 80.\",\"D\":\"Decomposition peaked at 40 earthworms per m², after which leaf mass lost declined as density increased further.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"A\"},{\"id\":\"sat99-rw-m1-q15\",\"number\":15,\"type\":\"data_analysis\",\"passage\":\"A team of behavioral economists led by Rajiv Chandra examined whether the timing of a charitable appeal influences donation amounts. Participants were randomly assigned to receive a solicitation email either on the morning of payday, three days after payday, or ten days after payday. The researchers recorded the mean donation in each condition as well as the percentage of recipients who donated at all.\",\"visual\":{\"type\":\"table\",\"title\":\"Donation Behavior by Email Timing\",\"headers\":[\"Timing of Appeal\",\"Donation Rate (%)\",\"Mean Donation ($)\"],\"rows\":[[\"Morning of payday\",\"34\",\"22.50\"],[\"3 days after payday\",\"29\",\"18.75\"],[\"10 days after payday\",\"18\",\"11.20\"]]},\"prompt\":\"Which choice most accurately describes the relationship between appeal timing and donor behavior shown in the table?\",\"choices\":{\"A\":\"As the delay after payday increased, the donation rate rose while the mean donation amount fell.\",\"B\":\"As the delay after payday increased, both the donation rate and the mean donation amount declined.\",\"C\":\"The donation rate was highest ten days after payday, but the mean donation was highest on the morning of payday.\",\"D\":\"Appeals sent three days after payday produced both the highest donation rate and the highest mean donation.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m1-q16\",\"number\":16,\"type\":\"data_analysis\",\"passage\":\"Materials scientist Inés Beltrán tested the tensile strength of a novel bio-based polymer at several curing temperatures to determine an optimal processing condition. She noted that insufficient curing leaves polymer chains weakly bonded, whereas excessive heat can degrade the chains and reduce strength. Beltrán predicted that tensile strength would rise with temperature up to a threshold and then fall beyond it.\",\"visual\":{\"type\":\"table\",\"title\":\"Tensile Strength of Bio-Based Polymer by Curing Temperature\",\"headers\":[\"Curing Temperature (°C)\",\"Tensile Strength (MPa)\"],\"rows\":[[\"100\",\"42\"],[\"120\",\"58\"],[\"140\",\"71\"],[\"160\",\"65\"],[\"180\",\"49\"]]},\"prompt\":\"Which choice best describes data from the table that support Beltrán's prediction?\",\"choices\":{\"A\":\"Tensile strength rose continuously from 100 °C to 180 °C, confirming that higher curing temperatures always improve the polymer.\",\"B\":\"Tensile strength was lowest at 100 °C and highest at 180 °C, indicating that thermal degradation did not occur within the tested range.\",\"C\":\"Tensile strength increased from 100 °C to 140 °C and then decreased at 160 °C and 180 °C, indicating a peak within the tested range.\",\"D\":\"Tensile strength remained constant between 120 °C and 160 °C, suggesting that curing temperature has little effect in that interval.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m1-q17\",\"number\":17,\"type\":\"data_analysis\",\"passage\":\"In a study of cognitive load during multitasking, psychologist Theo Abedi asked participants to complete a primary reading comprehension task while performing a secondary task of varying difficulty. He measured reading accuracy (percentage of comprehension questions answered correctly) and reaction time on the secondary task. Abedi wanted to determine whether a moderately demanding secondary task would disrupt primary-task performance less than either a very easy or a very difficult one.\",\"visual\":{\"type\":\"table\",\"title\":\"Performance by Secondary-Task Difficulty\",\"headers\":[\"Secondary-Task Difficulty\",\"Reading Accuracy (%)\",\"Secondary-Task Reaction Time (ms)\"],\"rows\":[[\"Very easy\",\"74\",\"310\"],[\"Moderate\",\"88\",\"425\"],[\"Difficult\",\"61\",\"690\"]]},\"prompt\":\"Which choice best describes data from the table that are relevant to Abedi's research question?\",\"choices\":{\"A\":\"Reading accuracy was lower in the difficult condition than in the very easy condition, showing that harder secondary tasks always impair primary-task performance.\",\"B\":\"Reading accuracy was higher in the moderate condition than in either the very easy or the difficult condition, consistent with the idea that a moderately demanding secondary task disrupts reading least.\",\"C\":\"Reaction time on the secondary task was shortest in the very easy condition, indicating that participants focused most effectively on reading when the secondary task was simplest.\",\"D\":\"Reading accuracy and secondary-task reaction time both increased as difficulty rose, suggesting that the two measures track each other across conditions.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"}]" }, { "costUsd": 0.150315, "model": "opus", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 4 SAT Reading & Writing questions.\nType: claim_support | Test: 99 | Unit: 1 | Start Q#: 18\n\nMANDATORY: Your JSON array MUST contain exactly 4 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nFollow the instructions for your type:\n\n[vocabulary_in_context]\n- 2-3 sentence passage with ______ blank from specific academic domains (molecular biology, Renaissance art, behavioral economics, astrophysics, political philosophy)\n- 4 college-level word choices from the same semantic field\n- Correct word is most PRECISE for context, not most common synonym\n- ID format: sat{testNumber}-rw-m{mod}-q{N}\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"vocabulary_in_context\",\"passage\":\"...with ______...\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[text_comprehension]\n- 3-8 sentence passage from published-quality source (journal, novel, essay) with specific names/dates/findings\n- Mix subtypes: main idea (2-3), author's purpose (1-2), inference (2-3), tone (0-1), structure (1-2)\n- Topics: natural science, social science, literature, history, arts/humanities (use each)\n- Distractors: surface-level misreading, wrong scope, opposite interpretation\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"text_comprehension\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[data_analysis]\n- Passage describing realistic data/study results requiring INTERPRETATION, not lookup\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Data patterns: diminishing returns, optimal values, conditional relationships\n- Distractors: wrong column, wrong trend direction, correlation vs causation confusion\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"data_analysis\",\"passage\":\"...\",\"visual\":{...},\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[claim_support]\n- Mix: \"Which finding supports/undermines the claim?\" AND \"Which quotation illustrates the claim?\"\n- Claims must be specific and testable. Distractors are topically related but don't address the specific mechanism.\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"claim_support\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[standard_english_conventions]\n- Passage with [underlined: text here] portion. Student picks best replacement.\n- Cover: subject-verb agreement, pronouns, verb tense, parallel structure, modifiers, commas, semicolons, apostrophes, sentence boundaries\n- \"NO CHANGE\" (choice A) correct ~25% of the time, NOT always\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"standard_english_conventions\",\"passage\":\"...with [underlined: ...]...\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[transitions]\n- Passage with ______ blank connecting two ideas. All 4 choices grammatically correct; distinction is logical.\n- Cover: contrast, addition, cause/effect, example, sequence, concession\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"transitions\",\"passage\":\"...______ ...\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[rhetorical_synthesis]\n- Notes format: \"While researching [topic], a student has taken these notes:\\n- Note 1\\n- Note 2\\n- Note 3\"\n- Specific rhetorical goal stated. All 4 choices are grammatical sentences using the notes; only 1 achieves the goal.\n- Goals: emphasize difference, highlight finding, introduce topic, summarize conclusion, present cause-effect\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"rhetorical_synthesis\",\"passage\":\"...\",\"prompt\":\"The student wants to [goal]. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\nOutput ONLY a valid JSON array of EXACTLY 4 objects. No markdown. Count them before outputting.", "text": "[{\"id\":\"sat99-rw-m1-q18\",\"number\":18,\"type\":\"claim_support\",\"passage\":\"A researcher studying the cognitive development of corvids (a family of birds including crows and ravens) has proposed that New Caledonian crows possess a form of causal reasoning that extends beyond simple trial-and-error learning. Specifically, the researcher claims that these birds can mentally simulate the physical consequences of an action before executing it, allowing them to solve novel tool-use problems on the first attempt without prior exposure.\",\"prompt\":\"Which finding, if true, would most directly support the researcher's claim?\",\"choices\":{\"A\":\"In laboratory settings, New Caledonian crows that observed other crows retrieving food with hooked sticks were subsequently able to use similar tools themselves after several practice sessions.\",\"B\":\"When presented with a novel apparatus requiring them to drop stones into a water-filled tube to raise a floating reward, naive New Caledonian crows selected stones of appropriate size and density on their first attempt, ignoring objects that would float.\",\"C\":\"New Caledonian crows in the wild have been documented manufacturing hooked tools from pandanus leaves, a behavior that varies regionally and appears to be transmitted across generations.\",\"D\":\"Neuroanatomical studies reveal that New Caledonian crows possess a higher density of neurons in the pallium than other corvid species of comparable body size.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m1-q19\",\"number\":19,\"type\":\"claim_support\",\"passage\":\"In her 2021 study of urban microclimates, geographer Lena Abernathy argues that the cooling effect of small neighborhood parks on surrounding city blocks is disproportionately large relative to their area—a phenomenon she attributes not to the parks' vegetation alone but to the way they disrupt the continuous surface of heat-absorbing pavement and buildings that characterize dense urban cores.\",\"prompt\":\"Which finding from Abernathy's study would most directly support her specific explanation for the cooling effect?\",\"choices\":{\"A\":\"Neighborhood parks measuring less than one hectare were associated with average temperature reductions of 1.8°C in adjacent blocks during summer afternoons.\",\"B\":\"Tree-canopy coverage within the parks correlated positively with the magnitude of temperature reduction observed in surrounding areas.\",\"C\":\"Paved plazas of equivalent size to the parks, when inserted into dense urban blocks, produced measurable cooling effects roughly 60 percent as large as those of vegetated parks.\",\"D\":\"Residents living within 200 meters of a neighborhood park reported lower rates of heat-related illness than residents living farther away during extreme heat events.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m1-q20\",\"number\":20,\"type\":\"claim_support\",\"passage\":\"In his critical study of the Victorian novel, literary scholar Marcus Holbrook contends that Elizabeth Gaskell's \\\"North and South\\\" (1855) departs from the conventions of the industrial novel by granting its female protagonist, Margaret Hale, a role as economic arbiter rather than as a passive moral influence. Holbrook argues that Margaret's decisions shape the material fortunes of the novel's central mill, a reversal of the typical Victorian pattern in which women inspire male action without directly determining business outcomes.\",\"prompt\":\"Which quotation from \\\"North and South\\\" would most effectively illustrate Holbrook's claim?\",\"choices\":{\"A\":\"\\\"Margaret felt, as she watched the crowds streaming from the mill gates, a sorrow she could scarcely name, a pity for the weary faces that seemed to reproach her own ease.\\\"\",\"B\":\"\\\"Mr. Thornton's manner toward her had grown gentler since their last conversation, and she was aware that her words had, in some measure, softened his views on the conduct of his workmen.\\\"\",\"C\":\"\\\"It was Margaret's capital, inherited from Mr. Bell and offered upon terms of her own devising, that permitted Thornton to reopen Marlborough Mills when no bank in Milton would advance him a shilling.\\\"\",\"D\":\"\\\"She thought often of the North, of its smoke and clangor, and wondered whether the gentleness of the South had not unfitted her for a world in which men like Thornton labored and contended.\\\"\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m1-q21\",\"number\":21,\"type\":\"claim_support\",\"passage\":\"Archaeologist Priya Chandran has argued that the residents of Çatalhöyük, a Neolithic settlement in southern Anatolia occupied from roughly 7100 to 5950 BCE, maintained a remarkably egalitarian social structure throughout the settlement's occupation, with no single household accumulating significantly greater wealth or status than its neighbors. Critics have suggested that this apparent egalitarianism may reflect the limitations of the archaeological record rather than the actual social organization of the community.\",\"prompt\":\"Which finding, if true, would most directly undermine Chandran's claim?\",\"choices\":{\"A\":\"Isotopic analysis of human remains from Çatalhöyük indicates that dietary composition varied little across individuals, suggesting broadly similar access to food resources.\",\"B\":\"Excavations at a contemporaneous settlement 200 kilometers to the east have revealed clear evidence of status differentiation, including elite burials with imported obsidian goods.\",\"C\":\"Detailed analysis of the distribution of rare imported materials across Çatalhöyük households reveals that a small cluster of dwellings consistently contained ten times the quantity found elsewhere in the settlement.\",\"D\":\"The layout of Çatalhöyük, with its densely packed houses sharing walls and accessed through roof openings, has been interpreted as evidence of strong communal identity.\",\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"}]" }, { "costUsd": 0.08124, "model": "opus", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 5 SAT Reading & Writing questions.\nType: standard_english_conventions | Test: 99 | Unit: 1 | Start Q#: 22\n\nMANDATORY: Your JSON array MUST contain exactly 5 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nFollow the instructions for your type:\n\n[vocabulary_in_context]\n- 2-3 sentence passage with ______ blank from specific academic domains (molecular biology, Renaissance art, behavioral economics, astrophysics, political philosophy)\n- 4 college-level word choices from the same semantic field\n- Correct word is most PRECISE for context, not most common synonym\n- ID format: sat{testNumber}-rw-m{mod}-q{N}\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"vocabulary_in_context\",\"passage\":\"...with ______...\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[text_comprehension]\n- 3-8 sentence passage from published-quality source (journal, novel, essay) with specific names/dates/findings\n- Mix subtypes: main idea (2-3), author's purpose (1-2), inference (2-3), tone (0-1), structure (1-2)\n- Topics: natural science, social science, literature, history, arts/humanities (use each)\n- Distractors: surface-level misreading, wrong scope, opposite interpretation\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"text_comprehension\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[data_analysis]\n- Passage describing realistic data/study results requiring INTERPRETATION, not lookup\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Data patterns: diminishing returns, optimal values, conditional relationships\n- Distractors: wrong column, wrong trend direction, correlation vs causation confusion\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"data_analysis\",\"passage\":\"...\",\"visual\":{...},\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[claim_support]\n- Mix: \"Which finding supports/undermines the claim?\" AND \"Which quotation illustrates the claim?\"\n- Claims must be specific and testable. Distractors are topically related but don't address the specific mechanism.\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"claim_support\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[standard_english_conventions]\n- Passage with [underlined: text here] portion. Student picks best replacement.\n- Cover: subject-verb agreement, pronouns, verb tense, parallel structure, modifiers, commas, semicolons, apostrophes, sentence boundaries\n- \"NO CHANGE\" (choice A) correct ~25% of the time, NOT always\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"standard_english_conventions\",\"passage\":\"...with [underlined: ...]...\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[transitions]\n- Passage with ______ blank connecting two ideas. All 4 choices grammatically correct; distinction is logical.\n- Cover: contrast, addition, cause/effect, example, sequence, concession\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"transitions\",\"passage\":\"...______ ...\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[rhetorical_synthesis]\n- Notes format: \"While researching [topic], a student has taken these notes:\\n- Note 1\\n- Note 2\\n- Note 3\"\n- Specific rhetorical goal stated. All 4 choices are grammatical sentences using the notes; only 1 achieves the goal.\n- Goals: emphasize difference, highlight finding, introduce topic, summarize conclusion, present cause-effect\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"rhetorical_synthesis\",\"passage\":\"...\",\"prompt\":\"The student wants to [goal]. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\nOutput ONLY a valid JSON array of EXACTLY 5 objects. No markdown. Count them before outputting.", "text": "[{\"id\":\"sat99-rw-m1-q22\",\"number\":22,\"type\":\"standard_english_conventions\",\"passage\":\"The sculptor Louise Nevelson, who emigrated from what is now Ukraine to the United States in 1905, became renowned for her monochromatic wooden assemblages; these works, often painted entirely black or white, [underlined: transforms] discarded objects into monumental compositions.\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"transform\",\"C\":\"is transforming\",\"D\":\"transforming\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m1-q23\",\"number\":23,\"type\":\"standard_english_conventions\",\"passage\":\"Marine biologist Sylvia Earle has spent decades advocating for ocean conservation. Her research expeditions have taken her to depths few humans have reached, and [underlined: they're findings have] reshaped public understanding of deep-sea ecosystems.\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"their findings have\",\"C\":\"there findings have\",\"D\":\"they're findings has\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m1-q24\",\"number\":24,\"type\":\"standard_english_conventions\",\"passage\":\"In traditional Japanese $kintsugi$, artisans repair broken pottery with lacquer mixed with powdered gold, silver, or [underlined: platinum, the technique treats] breakage and repair as part of an object's history rather than something to disguise.\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"platinum the technique treats\",\"C\":\"platinum; the technique treats\",\"D\":\"platinum, treating\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m1-q25\",\"number\":25,\"type\":\"standard_english_conventions\",\"passage\":\"Economist Esther Duflo, who shared the 2019 Nobel Prize for her experimental approach to alleviating global poverty, argues that small-scale randomized trials can reveal which interventions actually [underlined: works] in specific local contexts.\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"is working\",\"C\":\"has worked\",\"D\":\"work\"},\"answer_format\":\"mcq\",\"correct_answer\":\"D\"},{\"id\":\"sat99-rw-m1-q26\",\"number\":26,\"type\":\"standard_english_conventions\",\"passage\":\"By the time astronomers confirmed the existence of the first exoplanet orbiting a sun-like star in 1995, decades of speculation about worlds beyond our solar [underlined: system, had] given way to a concrete, observable field of research.\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"system had\",\"C\":\"system; had\",\"D\":\"system: had\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"}]" }, { "costUsd": 0.07734, "model": "opus", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 4 SAT Reading & Writing questions.\nType: transitions | Test: 99 | Unit: 1 | Start Q#: 27\n\nMANDATORY: Your JSON array MUST contain exactly 4 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nFollow the instructions for your type:\n\n[vocabulary_in_context]\n- 2-3 sentence passage with ______ blank from specific academic domains (molecular biology, Renaissance art, behavioral economics, astrophysics, political philosophy)\n- 4 college-level word choices from the same semantic field\n- Correct word is most PRECISE for context, not most common synonym\n- ID format: sat{testNumber}-rw-m{mod}-q{N}\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"vocabulary_in_context\",\"passage\":\"...with ______...\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[text_comprehension]\n- 3-8 sentence passage from published-quality source (journal, novel, essay) with specific names/dates/findings\n- Mix subtypes: main idea (2-3), author's purpose (1-2), inference (2-3), tone (0-1), structure (1-2)\n- Topics: natural science, social science, literature, history, arts/humanities (use each)\n- Distractors: surface-level misreading, wrong scope, opposite interpretation\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"text_comprehension\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[data_analysis]\n- Passage describing realistic data/study results requiring INTERPRETATION, not lookup\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Data patterns: diminishing returns, optimal values, conditional relationships\n- Distractors: wrong column, wrong trend direction, correlation vs causation confusion\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"data_analysis\",\"passage\":\"...\",\"visual\":{...},\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[claim_support]\n- Mix: \"Which finding supports/undermines the claim?\" AND \"Which quotation illustrates the claim?\"\n- Claims must be specific and testable. Distractors are topically related but don't address the specific mechanism.\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"claim_support\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[standard_english_conventions]\n- Passage with [underlined: text here] portion. Student picks best replacement.\n- Cover: subject-verb agreement, pronouns, verb tense, parallel structure, modifiers, commas, semicolons, apostrophes, sentence boundaries\n- \"NO CHANGE\" (choice A) correct ~25% of the time, NOT always\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"standard_english_conventions\",\"passage\":\"...with [underlined: ...]...\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[transitions]\n- Passage with ______ blank connecting two ideas. All 4 choices grammatically correct; distinction is logical.\n- Cover: contrast, addition, cause/effect, example, sequence, concession\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"transitions\",\"passage\":\"...______ ...\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[rhetorical_synthesis]\n- Notes format: \"While researching [topic], a student has taken these notes:\\n- Note 1\\n- Note 2\\n- Note 3\"\n- Specific rhetorical goal stated. All 4 choices are grammatical sentences using the notes; only 1 achieves the goal.\n- Goals: emphasize difference, highlight finding, introduce topic, summarize conclusion, present cause-effect\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"rhetorical_synthesis\",\"passage\":\"...\",\"prompt\":\"The student wants to [goal]. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\nOutput ONLY a valid JSON array of EXACTLY 4 objects. No markdown. Count them before outputting.", "text": "[{\"id\":\"sat99-rw-m1-q27\",\"number\":27,\"type\":\"transitions\",\"passage\":\"The axolotl, a salamander native to lakes near Mexico City, retains its larval features—including external gills and a finned tail—throughout its adult life, a phenomenon known as neoteny. Most salamanders undergo metamorphosis, shedding these juvenile traits as they mature. The axolotl, ______ remains perpetually youthful in appearance, a trait that has made it a valuable subject for studies of regeneration and development.\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"by contrast,\",\"B\":\"for instance,\",\"C\":\"as a result,\",\"D\":\"in addition,\"},\"answer_format\":\"mcq\",\"correct_answer\":\"A\"},{\"id\":\"sat99-rw-m1-q28\",\"number\":28,\"type\":\"transitions\",\"passage\":\"Archaeologist Sarah Parcak uses satellite imagery to identify buried archaeological sites, detecting subtle variations in soil and vegetation that hint at structures beneath the surface. Traditional excavation methods can take years and require substantial funding to survey even a small region. Parcak's technique, ______ allows researchers to map hundreds of potential sites in a fraction of the time, dramatically accelerating the pace of discovery.\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"nevertheless,\",\"B\":\"likewise,\",\"C\":\"however,\",\"D\":\"for example,\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m1-q29\",\"number\":29,\"type\":\"transitions\",\"passage\":\"Economist Esther Duflo and her colleagues tested whether small financial incentives could increase childhood immunization rates in rural Rajasthan. Villages offering a kilogram of lentils per vaccination saw immunization rates rise from 6 percent to 38 percent. ______ the program proved more cost-effective than conventional outreach, because the higher participation reduced the per-child expense of operating each mobile clinic.\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"Nonetheless,\",\"B\":\"Similarly,\",\"C\":\"Moreover,\",\"D\":\"Instead,\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m1-q30\",\"number\":30,\"type\":\"transitions\",\"passage\":\"In the late 1870s, painter Mary Cassatt was invited by Edgar Degas to exhibit with the Impressionists in Paris, a group then regarded by critics as radical and unpolished. Cassatt's decision to align herself with these artists cost her the patronage of several conservative American collectors. ______ the affiliation gave her access to a circle of experimental printmakers whose techniques would reshape her own approach to composition.\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"Likewise,\",\"B\":\"In turn,\",\"C\":\"Therefore,\",\"D\":\"Still,\"},\"answer_format\":\"mcq\",\"correct_answer\":\"D\"}]" }, { "costUsd": 0.121365, "model": "opus", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 3 SAT Reading & Writing questions.\nType: rhetorical_synthesis | Test: 99 | Unit: 1 | Start Q#: 31\n\nMANDATORY: Your JSON array MUST contain exactly 3 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nFollow the instructions for your type:\n\n[vocabulary_in_context]\n- 2-3 sentence passage with ______ blank from specific academic domains (molecular biology, Renaissance art, behavioral economics, astrophysics, political philosophy)\n- 4 college-level word choices from the same semantic field\n- Correct word is most PRECISE for context, not most common synonym\n- ID format: sat{testNumber}-rw-m{mod}-q{N}\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"vocabulary_in_context\",\"passage\":\"...with ______...\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[text_comprehension]\n- 3-8 sentence passage from published-quality source (journal, novel, essay) with specific names/dates/findings\n- Mix subtypes: main idea (2-3), author's purpose (1-2), inference (2-3), tone (0-1), structure (1-2)\n- Topics: natural science, social science, literature, history, arts/humanities (use each)\n- Distractors: surface-level misreading, wrong scope, opposite interpretation\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"text_comprehension\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[data_analysis]\n- Passage describing realistic data/study results requiring INTERPRETATION, not lookup\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Data patterns: diminishing returns, optimal values, conditional relationships\n- Distractors: wrong column, wrong trend direction, correlation vs causation confusion\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"data_analysis\",\"passage\":\"...\",\"visual\":{...},\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[claim_support]\n- Mix: \"Which finding supports/undermines the claim?\" AND \"Which quotation illustrates the claim?\"\n- Claims must be specific and testable. Distractors are topically related but don't address the specific mechanism.\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"claim_support\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[standard_english_conventions]\n- Passage with [underlined: text here] portion. Student picks best replacement.\n- Cover: subject-verb agreement, pronouns, verb tense, parallel structure, modifiers, commas, semicolons, apostrophes, sentence boundaries\n- \"NO CHANGE\" (choice A) correct ~25% of the time, NOT always\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"standard_english_conventions\",\"passage\":\"...with [underlined: ...]...\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[transitions]\n- Passage with ______ blank connecting two ideas. All 4 choices grammatically correct; distinction is logical.\n- Cover: contrast, addition, cause/effect, example, sequence, concession\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"transitions\",\"passage\":\"...______ ...\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[rhetorical_synthesis]\n- Notes format: \"While researching [topic], a student has taken these notes:\\n- Note 1\\n- Note 2\\n- Note 3\"\n- Specific rhetorical goal stated. All 4 choices are grammatical sentences using the notes; only 1 achieves the goal.\n- Goals: emphasize difference, highlight finding, introduce topic, summarize conclusion, present cause-effect\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"rhetorical_synthesis\",\"passage\":\"...\",\"prompt\":\"The student wants to [goal]. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\nOutput ONLY a valid JSON array of EXACTLY 3 objects. No markdown. Count them before outputting.", "text": "[{\"id\":\"sat99-rw-m1-q31\",\"number\":31,\"type\":\"rhetorical_synthesis\",\"passage\":\"While researching the migratory patterns of the Arctic tern, a student has taken these notes:\\n- The Arctic tern (Sterna paradisaea) is a small seabird with a wingspan of about 75 cm.\\n- It breeds in Arctic regions during the Northern Hemisphere summer.\\n- It winters in Antarctic waters during the Southern Hemisphere summer.\\n- A 2010 tracking study by Egevang et al. found terns travel roughly 71,000 km annually.\\n- This is the longest known migration of any animal on Earth.\",\"prompt\":\"The student wants to emphasize the remarkable nature of the Arctic tern's migration distance. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"The Arctic tern, a small seabird with a 75 cm wingspan, breeds in the Arctic and winters in Antarctic waters.\",\"B\":\"A 2010 tracking study by Egevang et al. revealed that the Arctic tern travels roughly 71,000 km annually—the longest known migration of any animal on Earth.\",\"C\":\"The Arctic tern breeds in Arctic regions during the Northern Hemisphere summer and migrates to Antarctic waters for the Southern Hemisphere summer.\",\"D\":\"Egevang and colleagues conducted a tracking study of the Arctic tern, a seabird whose scientific name is Sterna paradisaea, in 2010.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m1-q32\",\"number\":32,\"type\":\"rhetorical_synthesis\",\"passage\":\"While researching the painter Artemisia Gentileschi, a student has taken these notes:\\n- Artemisia Gentileschi (1593–1656) was an Italian Baroque painter.\\n- She trained in her father Orazio's workshop in Rome.\\n- In 1616, she became the first woman admitted to the Accademia delle Arti del Disegno in Florence.\\n- Her painting Judith Slaying Holofernes (c. 1620) is held by the Uffizi Gallery.\\n- Art historians cite the work for its dramatic chiaroscuro and unflinching depiction of violence.\",\"prompt\":\"The student wants to introduce Artemisia Gentileschi to an audience unfamiliar with her work. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"Judith Slaying Holofernes, painted around 1620, is noted by art historians for its dramatic chiaroscuro and unflinching depiction of violence.\",\"B\":\"In 1616, Artemisia Gentileschi became the first woman admitted to the Accademia delle Arti del Disegno in Florence.\",\"C\":\"Artemisia Gentileschi (1593–1656) was an Italian Baroque painter whose works, such as Judith Slaying Holofernes, are celebrated for their dramatic intensity.\",\"D\":\"Trained in her father Orazio's Roman workshop, Artemisia Gentileschi later moved to Florence, where the Uffizi Gallery now holds one of her paintings.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m1-q33\",\"number\":33,\"type\":\"rhetorical_synthesis\",\"passage\":\"While researching a behavioral economics experiment, a student has taken these notes:\\n- Researchers Dan Ariely and colleagues studied how anchoring affects willingness to pay.\\n- Participants were asked to write down the last two digits of their Social Security number.\\n- They were then asked to bid on items such as wine, chocolates, and electronics.\\n- Participants with higher digit pairs (80–99) bid up to 346% more than those with lower pairs (00–19).\\n- The digits had no logical connection to the items' actual value.\",\"prompt\":\"The student wants to highlight the central finding of the experiment. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"Dan Ariely and his colleagues conducted a behavioral economics experiment in which participants bid on items including wine, chocolates, and electronics.\",\"B\":\"Although the last two digits of a Social Security number bear no logical relation to an item's value, participants with higher digit pairs bid up to 346% more than those with lower pairs.\",\"C\":\"In one behavioral economics study, participants were asked to write down the last two digits of their Social Security number before bidding on a series of consumer items.\",\"D\":\"Anchoring is a cognitive bias in which an arbitrary number influences subsequent judgments, as Dan Ariely and colleagues demonstrated in their experiment.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"}]" }, { "costUsd": 0.08304, "model": "opus", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 5 SAT Reading & Writing questions.\nType: vocabulary_in_context | Test: 99 | Unit: 2 | Start Q#: 1\n\nMANDATORY: Your JSON array MUST contain exactly 5 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nFollow the instructions for your type:\n\n[vocabulary_in_context]\n- 2-3 sentence passage with ______ blank from specific academic domains (molecular biology, Renaissance art, behavioral economics, astrophysics, political philosophy)\n- 4 college-level word choices from the same semantic field\n- Correct word is most PRECISE for context, not most common synonym\n- ID format: sat{testNumber}-rw-m{mod}-q{N}\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"vocabulary_in_context\",\"passage\":\"...with ______...\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[text_comprehension]\n- 3-8 sentence passage from published-quality source (journal, novel, essay) with specific names/dates/findings\n- Mix subtypes: main idea (2-3), author's purpose (1-2), inference (2-3), tone (0-1), structure (1-2)\n- Topics: natural science, social science, literature, history, arts/humanities (use each)\n- Distractors: surface-level misreading, wrong scope, opposite interpretation\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"text_comprehension\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[data_analysis]\n- Passage describing realistic data/study results requiring INTERPRETATION, not lookup\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Data patterns: diminishing returns, optimal values, conditional relationships\n- Distractors: wrong column, wrong trend direction, correlation vs causation confusion\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"data_analysis\",\"passage\":\"...\",\"visual\":{...},\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[claim_support]\n- Mix: \"Which finding supports/undermines the claim?\" AND \"Which quotation illustrates the claim?\"\n- Claims must be specific and testable. Distractors are topically related but don't address the specific mechanism.\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"claim_support\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[standard_english_conventions]\n- Passage with [underlined: text here] portion. Student picks best replacement.\n- Cover: subject-verb agreement, pronouns, verb tense, parallel structure, modifiers, commas, semicolons, apostrophes, sentence boundaries\n- \"NO CHANGE\" (choice A) correct ~25% of the time, NOT always\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"standard_english_conventions\",\"passage\":\"...with [underlined: ...]...\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[transitions]\n- Passage with ______ blank connecting two ideas. All 4 choices grammatically correct; distinction is logical.\n- Cover: contrast, addition, cause/effect, example, sequence, concession\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"transitions\",\"passage\":\"...______ ...\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[rhetorical_synthesis]\n- Notes format: \"While researching [topic], a student has taken these notes:\\n- Note 1\\n- Note 2\\n- Note 3\"\n- Specific rhetorical goal stated. All 4 choices are grammatical sentences using the notes; only 1 achieves the goal.\n- Goals: emphasize difference, highlight finding, introduce topic, summarize conclusion, present cause-effect\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"rhetorical_synthesis\",\"passage\":\"...\",\"prompt\":\"The student wants to [goal]. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\nOutput ONLY a valid JSON array of EXACTLY 5 objects. No markdown. Count them before outputting.", "text": "[{\"id\":\"sat99-rw-m2-q1\",\"number\":1,\"type\":\"vocabulary_in_context\",\"passage\":\"In studies of chromatin remodeling, researchers have observed that certain ATP-dependent complexes do not merely displace nucleosomes but ______ them along the DNA strand, repositioning histone octamers with remarkable spatial precision.\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"translocate\",\"B\":\"transport\",\"C\":\"transfer\",\"D\":\"transmit\"},\"answer_format\":\"mcq\",\"correct_answer\":\"A\"},{\"id\":\"sat99-rw-m2-q2\",\"number\":2,\"type\":\"vocabulary_in_context\",\"passage\":\"Art historian Patricia Rubin has argued that Vasari's biographies of Renaissance painters were never intended as neutral records; rather, they ______ a Florentine-centered narrative in which artists outside the Medici orbit were systematically diminished. This editorial shaping, she contends, distorted the historiography of Italian art for centuries.\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"disclosed\",\"B\":\"advanced\",\"C\":\"tolerated\",\"D\":\"inherited\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m2-q3\",\"number\":3,\"type\":\"vocabulary_in_context\",\"passage\":\"In behavioral economics, Richard Thaler's concept of mental accounting suggests that consumers do not treat money as fully ______. A tax refund, a birthday gift, and a paycheck of equal value will be spent quite differently, even though rational choice theory predicts identical treatment.\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"divisible\",\"B\":\"transparent\",\"C\":\"fungible\",\"D\":\"negotiable\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m2-q4\",\"number\":4,\"type\":\"vocabulary_in_context\",\"passage\":\"Observations from the James Webb Space Telescope have revealed galaxies at redshifts exceeding $z=13$ that appear far more luminous and structurally mature than existing cosmological models ______. Theorists are now reconsidering whether early star formation proceeded at rates previously thought impossible.\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"accommodate\",\"B\":\"recognize\",\"C\":\"endorse\",\"D\":\"withstand\"},\"answer_format\":\"mcq\",\"correct_answer\":\"A\"},{\"id\":\"sat99-rw-m2-q5\",\"number\":5,\"type\":\"vocabulary_in_context\",\"passage\":\"In \\\"On Liberty,\\\" John Stuart Mill warns that a citizenry may grow so accustomed to the tyranny of majority opinion that dissent becomes not merely unwelcome but nearly ______. Social pressure, he argued, can silence heterodox thought more effectively than any formal censor.\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"unthinkable\",\"B\":\"indifferent\",\"C\":\"redundant\",\"D\":\"ambivalent\"},\"answer_format\":\"mcq\",\"correct_answer\":\"A\"}]" }, { "costUsd": 0.272115, "model": "opus", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 8 SAT Reading & Writing questions.\nType: text_comprehension | Test: 99 | Unit: 2 | Start Q#: 6\n\nMANDATORY: Your JSON array MUST contain exactly 8 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nFollow the instructions for your type:\n\n[vocabulary_in_context]\n- 2-3 sentence passage with ______ blank from specific academic domains (molecular biology, Renaissance art, behavioral economics, astrophysics, political philosophy)\n- 4 college-level word choices from the same semantic field\n- Correct word is most PRECISE for context, not most common synonym\n- ID format: sat{testNumber}-rw-m{mod}-q{N}\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"vocabulary_in_context\",\"passage\":\"...with ______...\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[text_comprehension]\n- 3-8 sentence passage from published-quality source (journal, novel, essay) with specific names/dates/findings\n- Mix subtypes: main idea (2-3), author's purpose (1-2), inference (2-3), tone (0-1), structure (1-2)\n- Topics: natural science, social science, literature, history, arts/humanities (use each)\n- Distractors: surface-level misreading, wrong scope, opposite interpretation\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"text_comprehension\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[data_analysis]\n- Passage describing realistic data/study results requiring INTERPRETATION, not lookup\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Data patterns: diminishing returns, optimal values, conditional relationships\n- Distractors: wrong column, wrong trend direction, correlation vs causation confusion\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"data_analysis\",\"passage\":\"...\",\"visual\":{...},\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[claim_support]\n- Mix: \"Which finding supports/undermines the claim?\" AND \"Which quotation illustrates the claim?\"\n- Claims must be specific and testable. Distractors are topically related but don't address the specific mechanism.\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"claim_support\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[standard_english_conventions]\n- Passage with [underlined: text here] portion. Student picks best replacement.\n- Cover: subject-verb agreement, pronouns, verb tense, parallel structure, modifiers, commas, semicolons, apostrophes, sentence boundaries\n- \"NO CHANGE\" (choice A) correct ~25% of the time, NOT always\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"standard_english_conventions\",\"passage\":\"...with [underlined: ...]...\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[transitions]\n- Passage with ______ blank connecting two ideas. All 4 choices grammatically correct; distinction is logical.\n- Cover: contrast, addition, cause/effect, example, sequence, concession\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"transitions\",\"passage\":\"...______ ...\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[rhetorical_synthesis]\n- Notes format: \"While researching [topic], a student has taken these notes:\\n- Note 1\\n- Note 2\\n- Note 3\"\n- Specific rhetorical goal stated. All 4 choices are grammatical sentences using the notes; only 1 achieves the goal.\n- Goals: emphasize difference, highlight finding, introduce topic, summarize conclusion, present cause-effect\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"rhetorical_synthesis\",\"passage\":\"...\",\"prompt\":\"The student wants to [goal]. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\nOutput ONLY a valid JSON array of EXACTLY 8 objects. No markdown. Count them before outputting.", "text": "[{\"id\":\"sat99-rw-m2-q6\",\"number\":6,\"type\":\"text_comprehension\",\"passage\":\"In her 2019 study of medieval manuscript illumination, art historian Elina Gertsman challenges the long-standing assumption that blank spaces in devotional texts were merely unfinished work awaiting decoration. Drawing on a corpus of over two hundred Books of Hours, Gertsman argues that voids—empty medallions, unpainted margins, and unfilled initials—were often deliberate theological statements. The absence of image, she contends, invited contemplative engagement with the unrepresentable nature of the divine, functioning as a visual corollary to apophatic theology.\",\"prompt\":\"Which choice best describes the main idea of the text?\",\"choices\":{\"A\":\"Gertsman's analysis of Books of Hours reveals that medieval illuminators frequently left manuscripts unfinished due to economic pressures from wealthy patrons.\",\"B\":\"Gertsman proposes that empty spaces in medieval devotional manuscripts should be understood as intentional theological expressions rather than incomplete craftsmanship.\",\"C\":\"Gertsman demonstrates that apophatic theology originated in the artistic practices of medieval manuscript illuminators rather than in formal philosophical treatises.\",\"D\":\"Gertsman's corpus of two hundred Books of Hours proves that medieval viewers preferred unadorned manuscripts to heavily illuminated ones.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m2-q7\",\"number\":7,\"type\":\"text_comprehension\",\"passage\":\"The following is adapted from Willa Cather's 1918 novel My Ántonia. The narrator, Jim Burden, recalls his arrival on the Nebraska prairie as a child.\\n\\nI had the feeling that the world was left behind, that we had got over the edge of it, and were outside man's jurisdiction. I had never before looked up at the sky when there was not a familiar mountain ridge against it. But this was the complete dome of heaven, all there was of it. Between that earth and that sky I felt erased, blotted out. I did not say my prayers that night: here, I felt, what would be would be.\",\"prompt\":\"Which choice best describes the function of the underlined sentence \\\"Between that earth and that sky I felt erased, blotted out\\\" in the text as a whole?\",\"choices\":{\"A\":\"It introduces a geographic description that the narrator will later contradict with more detailed observations.\",\"B\":\"It conveys the narrator's fear that the prairie's inhabitants will not welcome a newcomer from the East.\",\"C\":\"It captures a sense of self-dissolution that motivates the narrator's subsequent abandonment of a familiar ritual.\",\"D\":\"It criticizes the narrator's earlier assumption that mountains were necessary for a meaningful landscape.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m2-q8\",\"number\":8,\"type\":\"text_comprehension\",\"passage\":\"Economist Esther Duflo's randomized controlled trials in rural Rajasthan produced a counterintuitive finding: offering mothers a small bag of lentils as an incentive for vaccinating their children raised full-immunization rates from 6 percent to 39 percent, while the program's per-child cost actually fell. The lentils, Duflo observed, did not persuade skeptics of vaccine safety; rather, they solved a subtler problem. Many parents already intended to immunize but perpetually postponed the clinic visit. The modest reward supplied the nudge needed to convert latent intention into action.\",\"prompt\":\"Based on the text, what can reasonably be inferred about the parents in Duflo's study?\",\"choices\":{\"A\":\"Most parents were skeptical of vaccine safety and required material incentives to overcome their doubts.\",\"B\":\"Parents in rural Rajasthan generally placed greater value on food staples than on their children's long-term health.\",\"C\":\"Many parents' failure to vaccinate reflected procrastination rather than an objection to immunization itself.\",\"D\":\"The program's success depended primarily on reducing the travel distance between villages and immunization clinics.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m2-q9\",\"number\":9,\"type\":\"text_comprehension\",\"passage\":\"Paleontologist Jingmai O'Connor has argued that the traditional dichotomy between \\\"avian\\\" and \\\"non-avian\\\" dinosaurs obscures more than it clarifies. Her analyses of Cretaceous specimens from the Jehol Biota suggest that features once considered diagnostic of birds—feathers, wishbones, even rudimentary flight—appeared sporadically across multiple theropod lineages. To speak of \\\"the origin of birds,\\\" O'Connor writes, is to impose a retrospective clarity on what was, in the Mesozoic, a thicket of parallel experiments.\",\"prompt\":\"Which choice best describes the author's purpose in the text?\",\"choices\":{\"A\":\"To argue that O'Connor's research has definitively resolved longstanding debates about which theropod lineage gave rise to modern birds.\",\"B\":\"To explain how O'Connor's work complicates a familiar evolutionary narrative by emphasizing the distributed nature of bird-like traits.\",\"C\":\"To criticize paleontologists who continue to use the terms \\\"avian\\\" and \\\"non-avian\\\" despite evidence from the Jehol Biota.\",\"D\":\"To describe a methodological controversy about how fossil specimens from the Cretaceous should be classified and catalogued.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m2-q10\",\"number\":10,\"type\":\"text_comprehension\",\"passage\":\"In a 2021 review of urban heat-island mitigation, civil engineer Hashem Akbari surveyed three decades of rooftop albedo studies. He found that while \\\"cool roofs\\\" painted with highly reflective coatings reliably reduced indoor cooling loads by 10–20 percent in low-rise buildings, their neighborhood-scale effects were far less consistent. In some Mediterranean cities, reflected solar radiation from dense clusters of cool roofs modestly raised street-level temperatures for pedestrians, particularly in narrow canyons where the reflected light was repeatedly absorbed by adjacent walls.\",\"prompt\":\"Which choice best describes the structure of the text?\",\"choices\":{\"A\":\"It presents a widely accepted engineering solution and then describes the mechanism by which it was originally developed.\",\"B\":\"It contrasts two competing theories about urban heat mitigation and argues in favor of the more recent one.\",\"C\":\"It introduces a well-established benefit of a technology and then qualifies that benefit by noting a context-dependent drawback.\",\"D\":\"It summarizes a historical debate and concludes that the debate has been resolved by Akbari's 2021 review.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m2-q11\",\"number\":11,\"type\":\"text_comprehension\",\"passage\":\"The following is adapted from an 1891 essay by Vernon Lee on the aesthetics of music.\\n\\nI do not deny that a sonata may be dissected, its modulations charted, its thematic transformations labeled with scholarly precision. Such labor has its uses. But let us not pretend that the sum of these annotations is the sonata. The listener who emerges from the concert hall transfigured has not been moved by a catalogue of devices; she has been moved by something the catalogue cannot capture, and which the catalogue was never designed to record.\",\"prompt\":\"Which choice best describes the tone of the passage?\",\"choices\":{\"A\":\"Dismissive—Lee rejects musical analysis as a futile endeavor that produces no genuine understanding.\",\"B\":\"Measured—Lee concedes the value of technical analysis while insisting on its insufficiency for explaining musical experience.\",\"C\":\"Nostalgic—Lee laments the decline of an earlier era in which listeners were more emotionally responsive to music.\",\"D\":\"Defensive—Lee justifies her own lack of formal musical training by attacking those who possess it.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m2-q12\",\"number\":12,\"type\":\"text_comprehension\",\"passage\":\"Historians of the early Qing dynasty have long debated the significance of the 1645 queue edict, which required Han Chinese men to adopt the Manchu hairstyle on pain of death. Recent work by Joanna Waley-Cohen reframes the edict as less an assertion of ethnic supremacy than a calculated test of political loyalty. Because the hairstyle was visible at a glance, it allowed Qing officials to identify dissent without the laborious inquisitions that had undermined earlier dynasties. The edict's cruelty, Waley-Cohen suggests, should not obscure its administrative elegance.\",\"prompt\":\"Based on the text, which of the following best describes Waley-Cohen's interpretation of the queue edict?\",\"choices\":{\"A\":\"It was primarily an expression of Manchu cultural chauvinism against Han Chinese traditions.\",\"B\":\"It was a comparatively humane policy when measured against the inquisitions of earlier dynasties.\",\"C\":\"It functioned as an efficient surveillance mechanism that reduced the need for more invasive forms of investigation.\",\"D\":\"It was an impractical policy whose administrative costs outweighed its benefits to the Qing state.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m2-q13\",\"number\":13,\"type\":\"text_comprehension\",\"passage\":\"When the astronomer Vera Rubin began measuring the rotation curves of spiral galaxies in the 1970s, she expected the outer stars to orbit more slowly than those near the galactic center, just as distant planets orbit the sun more slowly than nearby ones. Instead, her spectra showed that stars at the edges of galaxies like Andromeda were moving just as fast as those much closer in. The implication, Rubin recognized, was either that the law of gravity behaved differently on galactic scales or that each galaxy was embedded in a vast halo of unseen mass.\",\"prompt\":\"Which choice best states the main idea of the text?\",\"choices\":{\"A\":\"Rubin's spectra of Andromeda disproved the Newtonian laws of gravity and established a new framework for galactic dynamics.\",\"B\":\"Rubin's observations of galactic rotation contradicted her expectations and pointed toward two possible explanations with profound implications.\",\"C\":\"Rubin's work demonstrated that spiral galaxies rotate differently from solar systems because their stars are more massive than planets.\",\"D\":\"Rubin initially misinterpreted her spectra but later corrected her findings after comparing them with planetary orbital data.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"}]" }, { "costUsd": 0.17019, "model": "opus", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 4 SAT Reading & Writing questions.\nType: data_analysis | Test: 99 | Unit: 2 | Start Q#: 14\n\nMANDATORY: Your JSON array MUST contain exactly 4 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nFollow the instructions for your type:\n\n[vocabulary_in_context]\n- 2-3 sentence passage with ______ blank from specific academic domains (molecular biology, Renaissance art, behavioral economics, astrophysics, political philosophy)\n- 4 college-level word choices from the same semantic field\n- Correct word is most PRECISE for context, not most common synonym\n- ID format: sat{testNumber}-rw-m{mod}-q{N}\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"vocabulary_in_context\",\"passage\":\"...with ______...\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[text_comprehension]\n- 3-8 sentence passage from published-quality source (journal, novel, essay) with specific names/dates/findings\n- Mix subtypes: main idea (2-3), author's purpose (1-2), inference (2-3), tone (0-1), structure (1-2)\n- Topics: natural science, social science, literature, history, arts/humanities (use each)\n- Distractors: surface-level misreading, wrong scope, opposite interpretation\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"text_comprehension\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[data_analysis]\n- Passage describing realistic data/study results requiring INTERPRETATION, not lookup\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Data patterns: diminishing returns, optimal values, conditional relationships\n- Distractors: wrong column, wrong trend direction, correlation vs causation confusion\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"data_analysis\",\"passage\":\"...\",\"visual\":{...},\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[claim_support]\n- Mix: \"Which finding supports/undermines the claim?\" AND \"Which quotation illustrates the claim?\"\n- Claims must be specific and testable. Distractors are topically related but don't address the specific mechanism.\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"claim_support\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[standard_english_conventions]\n- Passage with [underlined: text here] portion. Student picks best replacement.\n- Cover: subject-verb agreement, pronouns, verb tense, parallel structure, modifiers, commas, semicolons, apostrophes, sentence boundaries\n- \"NO CHANGE\" (choice A) correct ~25% of the time, NOT always\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"standard_english_conventions\",\"passage\":\"...with [underlined: ...]...\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[transitions]\n- Passage with ______ blank connecting two ideas. All 4 choices grammatically correct; distinction is logical.\n- Cover: contrast, addition, cause/effect, example, sequence, concession\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"transitions\",\"passage\":\"...______ ...\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[rhetorical_synthesis]\n- Notes format: \"While researching [topic], a student has taken these notes:\\n- Note 1\\n- Note 2\\n- Note 3\"\n- Specific rhetorical goal stated. All 4 choices are grammatical sentences using the notes; only 1 achieves the goal.\n- Goals: emphasize difference, highlight finding, introduce topic, summarize conclusion, present cause-effect\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"rhetorical_synthesis\",\"passage\":\"...\",\"prompt\":\"The student wants to [goal]. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\nOutput ONLY a valid JSON array of EXACTLY 4 objects. No markdown. Count them before outputting.", "text": "[{\"id\":\"sat99-rw-m2-q14\",\"number\":14,\"type\":\"data_analysis\",\"passage\":\"Soil ecologist Marisol Vega investigated whether mycorrhizal fungi inoculation affects the drought tolerance of seedlings in degraded pastureland. She grew seedlings of four native tree species under identical water-restricted conditions, half with inoculation and half without, and measured survival rates after 90 days. Vega hypothesized that the benefit of inoculation would be greatest for species with naturally low drought tolerance, as measured by baseline (non-inoculated) survival.\",\"visual\":{\"type\":\"table\",\"title\":\"90-Day Seedling Survival Rates Under Drought Stress (%)\",\"headers\":[\"Species\",\"Non-inoculated\",\"Inoculated\",\"Change (pp)\"],\"rows\":[[\"Quercus rugosa\",\"72\",\"81\",\"+9\"],[\"Pinus leiophylla\",\"58\",\"74\",\"+16\"],[\"Alnus acuminata\",\"34\",\"62\",\"+28\"],[\"Prunus serotina\",\"49\",\"65\",\"+16\"]]},\"prompt\":\"Which choice best describes data from the table that support Vega's hypothesis?\",\"choices\":{\"A\":\"Alnus acuminata had the lowest non-inoculated survival rate and showed the largest increase in survival following inoculation.\",\"B\":\"Quercus rugosa had the highest non-inoculated survival rate and showed the largest increase in survival following inoculation.\",\"C\":\"Pinus leiophylla and Prunus serotina had identical increases in survival despite different baseline survival rates.\",\"D\":\"Every species showed an increase in survival following inoculation, with the average increase exceeding 15 percentage points.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"A\"},{\"id\":\"sat99-rw-m2-q15\",\"number\":15,\"type\":\"data_analysis\",\"passage\":\"In a study of consumer decision-making, economist Tarun Shah presented participants with a choice between two retirement savings plans that differed only in how contribution rates were framed: Plan X described contributions as a percentage of gross salary, while Plan Y described the same contributions as a fixed dollar amount per paycheck. Shah recorded the proportion of participants in each age group who selected Plan Y.\",\"visual\":{\"type\":\"table\",\"title\":\"Proportion Selecting Plan Y (Fixed-Dollar Framing) by Age Group\",\"headers\":[\"Age Group\",\"Plan Y Selection Rate\",\"Reported Confidence in Choice\"],\"rows\":[[\"18–29\",\"0.71\",\"Low\"],[\"30–44\",\"0.58\",\"Moderate\"],[\"45–59\",\"0.44\",\"Moderate\"],[\"60+\",\"0.29\",\"High\"]]},\"prompt\":\"Based on the table, which statement most accurately characterizes the relationship between age and framing preference?\",\"choices\":{\"A\":\"Older participants were more likely to select the fixed-dollar plan, suggesting increased susceptibility to concrete framing with age.\",\"B\":\"As age increased, preference for the fixed-dollar framing decreased, and reported confidence in the choice increased.\",\"C\":\"Confidence in the choice was lowest among participants who most frequently selected the percentage-based framing.\",\"D\":\"Age had no systematic effect on framing preference, since selection rates varied unpredictably across groups.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m2-q16\",\"number\":16,\"type\":\"data_analysis\",\"passage\":\"To assess whether urban noise affects the vocal behavior of songbirds, ornithologist Kenji Arai recorded dawn choruses of white-crowned sparrows at four sites in and around San Francisco. At each site, he measured the ambient noise level (in decibels) at 5:00 a.m. and the minimum frequency (in hertz) of the sparrows' songs, reasoning that birds in noisier environments might raise their minimum song frequency to avoid masking by low-frequency traffic noise.\",\"visual\":{\"type\":\"table\",\"title\":\"Ambient Noise and Minimum Song Frequency at Four Sites\",\"headers\":[\"Site\",\"Ambient Noise (dB)\",\"Minimum Song Frequency (Hz)\"],\"rows\":[[\"Presidio (suburban edge)\",\"42\",\"2,050\"],[\"Golden Gate Park\",\"51\",\"2,310\"],[\"Mission District\",\"63\",\"2,740\"],[\"SOMA (downtown)\",\"71\",\"3,020\"]]},\"prompt\":\"Which choice best describes data from the table that support Arai's reasoning?\",\"choices\":{\"A\":\"The Presidio site had the lowest ambient noise and the lowest minimum song frequency, while SOMA had the highest values for both.\",\"B\":\"Minimum song frequency at each site was more than 40 times the ambient noise level measured at that site.\",\"C\":\"The difference in minimum song frequency between Golden Gate Park and the Mission District was greater than the difference between the Presidio and Golden Gate Park.\",\"D\":\"Sites with higher ambient noise showed lower minimum song frequencies, contrary to what would be expected if traffic noise masked low-frequency song.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"A\"},{\"id\":\"sat99-rw-m2-q17\",\"number\":17,\"type\":\"data_analysis\",\"passage\":\"Historian Elena Costa examined the circulation numbers of four political pamphlets published in Venice between 1575 and 1590 to investigate whether length influenced popular reach. She hypothesized that shorter pamphlets would achieve wider circulation because they were cheaper to print and easier for semi-literate readers to consume in a single sitting.\",\"visual\":{\"type\":\"table\",\"title\":\"Venetian Political Pamphlets, 1575–1590\",\"headers\":[\"Pamphlet\",\"Length (pages)\",\"Estimated Copies Circulated\",\"Price (soldi)\"],\"rows\":[[\"Discorso sulla Repubblica\",\"48\",\"1,200\",\"12\"],[\"Lettera ai Cittadini\",\"8\",\"9,400\",\"2\"],[\"Trattato della Giustizia\",\"64\",\"850\",\"16\"],[\"Avviso al Popolo\",\"12\",\"7,100\",\"3\"]]},\"prompt\":\"Which choice best describes data from the table that support Costa's hypothesis?\",\"choices\":{\"A\":\"The two longest pamphlets, each exceeding 40 pages, circulated in fewer than 1,500 copies, while the two shortest, under 15 pages, each circulated in more than 7,000 copies.\",\"B\":\"Lettera ai Cittadini circulated in more copies than any other pamphlet, and its price was the lowest of the four.\",\"C\":\"The pamphlet with the highest price, Trattato della Giustizia, circulated in the fewest copies, suggesting that cost rather than length drove readership.\",\"D\":\"Avviso al Popolo was longer than Lettera ai Cittadini but circulated in fewer copies, showing that length had only a modest effect on reach.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"A\"}]" }, { "costUsd": 0.145665, "model": "opus", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 4 SAT Reading & Writing questions.\nType: claim_support | Test: 99 | Unit: 2 | Start Q#: 18\n\nMANDATORY: Your JSON array MUST contain exactly 4 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nFollow the instructions for your type:\n\n[vocabulary_in_context]\n- 2-3 sentence passage with ______ blank from specific academic domains (molecular biology, Renaissance art, behavioral economics, astrophysics, political philosophy)\n- 4 college-level word choices from the same semantic field\n- Correct word is most PRECISE for context, not most common synonym\n- ID format: sat{testNumber}-rw-m{mod}-q{N}\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"vocabulary_in_context\",\"passage\":\"...with ______...\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[text_comprehension]\n- 3-8 sentence passage from published-quality source (journal, novel, essay) with specific names/dates/findings\n- Mix subtypes: main idea (2-3), author's purpose (1-2), inference (2-3), tone (0-1), structure (1-2)\n- Topics: natural science, social science, literature, history, arts/humanities (use each)\n- Distractors: surface-level misreading, wrong scope, opposite interpretation\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"text_comprehension\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[data_analysis]\n- Passage describing realistic data/study results requiring INTERPRETATION, not lookup\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Data patterns: diminishing returns, optimal values, conditional relationships\n- Distractors: wrong column, wrong trend direction, correlation vs causation confusion\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"data_analysis\",\"passage\":\"...\",\"visual\":{...},\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[claim_support]\n- Mix: \"Which finding supports/undermines the claim?\" AND \"Which quotation illustrates the claim?\"\n- Claims must be specific and testable. Distractors are topically related but don't address the specific mechanism.\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"claim_support\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[standard_english_conventions]\n- Passage with [underlined: text here] portion. Student picks best replacement.\n- Cover: subject-verb agreement, pronouns, verb tense, parallel structure, modifiers, commas, semicolons, apostrophes, sentence boundaries\n- \"NO CHANGE\" (choice A) correct ~25% of the time, NOT always\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"standard_english_conventions\",\"passage\":\"...with [underlined: ...]...\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[transitions]\n- Passage with ______ blank connecting two ideas. All 4 choices grammatically correct; distinction is logical.\n- Cover: contrast, addition, cause/effect, example, sequence, concession\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"transitions\",\"passage\":\"...______ ...\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[rhetorical_synthesis]\n- Notes format: \"While researching [topic], a student has taken these notes:\\n- Note 1\\n- Note 2\\n- Note 3\"\n- Specific rhetorical goal stated. All 4 choices are grammatical sentences using the notes; only 1 achieves the goal.\n- Goals: emphasize difference, highlight finding, introduce topic, summarize conclusion, present cause-effect\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"rhetorical_synthesis\",\"passage\":\"...\",\"prompt\":\"The student wants to [goal]. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\nOutput ONLY a valid JSON array of EXACTLY 4 objects. No markdown. Count them before outputting.", "text": "[{\"id\":\"sat99-rw-m2-q18\",\"number\":18,\"type\":\"claim_support\",\"passage\":\"In a 2019 study, marine biologist Elena Vasquez examined the foraging behavior of Caribbean reef octopuses (Octopus briareus) and proposed that these cephalopods develop individualized hunting strategies that persist across multiple foraging trips, rather than responding identically to similar environmental stimuli. Vasquez argued that this consistency reflects stable behavioral differences between individuals analogous to personality traits in vertebrates.\",\"prompt\":\"Which finding from Vasquez's study, if true, would most directly support her claim?\",\"choices\":{\"A\":\"Octopuses in the study showed a marked preference for crustacean prey over mollusks, regardless of which prey type was more abundant in their immediate habitat.\",\"B\":\"When presented with identical artificial reef environments, individual octopuses each used distinct hunting techniques—such as ambush, pursuit, or den-baiting—that each animal reliably repeated across twelve separate trials.\",\"C\":\"Octopuses observed in the wild exhibited more varied hunting behaviors than those kept in controlled laboratory conditions, suggesting that environmental complexity shapes foraging choices.\",\"D\":\"Juvenile octopuses in the study displayed a wider range of hunting techniques than mature adults, whose behaviors became increasingly uniform as they aged.\",\"E\":\"\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m2-q19\",\"number\":19,\"type\":\"claim_support\",\"passage\":\"In her 2021 analysis of Dutch Golden Age still life painting, art historian Margriet de Vries contends that Clara Peeters's signature technique of embedding tiny self-portraits within the reflective surfaces of depicted pewter vessels was not merely a display of technical virtuosity but a deliberate assertion of authorial presence in a genre that typically effaced the painter's identity. De Vries argues that these reflections functioned as visual signatures at a time when female artists were frequently denied formal attribution.\",\"prompt\":\"Which finding, if true, would most strongly support de Vries's claim?\",\"choices\":{\"A\":\"X-ray analysis of several Peeters paintings reveals that the reflected self-portraits were painted in a final, separate layer using pigments distinct from those used in the surrounding still life composition.\",\"B\":\"Peeters's contemporaries, including male still life painters such as Osias Beert, occasionally included reflections of windows or studio interiors in their depicted metalware.\",\"C\":\"Archival records indicate that several of Peeters's still lifes were sold to wealthy patrons in Antwerp without any written attribution to the artist, and that she signed far fewer works than her male contemporaries.\",\"D\":\"Technical examination of pewter vessels from the early seventeenth century shows that their polished surfaces would have produced recognizable mirror-like reflections under studio lighting.\",\"E\":\"\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m2-q20\",\"number\":20,\"type\":\"claim_support\",\"passage\":\"In The Theory of Moral Sentiments (1759), Adam Smith argues that human beings possess an innate capacity for sympathy, which he defines not as pity but as the imaginative act of placing oneself in another's situation. Smith further claims that our moral judgments of others arise primarily from this sympathetic projection rather than from abstract reasoning about duty or consequence.\",\"prompt\":\"Which quotation from The Theory of Moral Sentiments most directly illustrates Smith's claim about the origin of moral judgment?\",\"choices\":{\"A\":\"\\\"Mankind, though naturally sympathetic, never conceive, for what has befallen another, that degree of passion which naturally animates the person principally concerned.\\\"\",\"B\":\"\\\"The man of the most perfect virtue... is he who joins, to the most perfect command of his own original and selfish feelings, the most exquisite sensibility both to the original and sympathetic feelings of others.\\\"\",\"C\":\"\\\"As we have no immediate experience of what other men feel, we can form no idea of the manner in which they are affected, but by conceiving what we ourselves should feel in the like situation.\\\"\",\"D\":\"\\\"Every faculty in one man is the measure by which he judges of the like faculty in another. I judge of your sight by my sight, of your ear by my ear, of your reason by my reason.\\\"\",\"E\":\"\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m2-q21\",\"number\":21,\"type\":\"claim_support\",\"passage\":\"Behavioral economist Priya Narayanan has proposed that so-called \\\"default effects\\\"—the tendency of people to stick with pre-selected options—stem not from genuine preference alignment but from the cognitive effort required to evaluate alternatives. Narayanan predicts that when decision-makers are given more time and cognitive resources to deliberate, the influence of defaults on their final choices will diminish substantially.\",\"prompt\":\"Which finding from a controlled experiment would most directly undermine Narayanan's claim?\",\"choices\":{\"A\":\"Participants who were given a forty-five minute deliberation period and reminded to consider all options chose the default option at essentially the same rate as participants who were required to decide within thirty seconds.\",\"B\":\"Participants under time pressure were more likely to report feeling anxious about their decisions than participants given ample time to deliberate, though both groups selected the default at similar rates.\",\"C\":\"Participants who were required to justify their decisions in writing chose the default less often than those who were not asked to provide justifications, regardless of how much time they were given.\",\"D\":\"Participants who were explicitly told which option had been pre-selected as the default were more likely to choose it than participants who were not informed about the default's status.\",\"E\":\"\"},\"answer_format\":\"mcq\",\"correct_answer\":\"A\"}]" }, { "costUsd": 0.07389, "model": "opus", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 5 SAT Reading & Writing questions.\nType: standard_english_conventions | Test: 99 | Unit: 2 | Start Q#: 22\n\nMANDATORY: Your JSON array MUST contain exactly 5 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nFollow the instructions for your type:\n\n[vocabulary_in_context]\n- 2-3 sentence passage with ______ blank from specific academic domains (molecular biology, Renaissance art, behavioral economics, astrophysics, political philosophy)\n- 4 college-level word choices from the same semantic field\n- Correct word is most PRECISE for context, not most common synonym\n- ID format: sat{testNumber}-rw-m{mod}-q{N}\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"vocabulary_in_context\",\"passage\":\"...with ______...\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[text_comprehension]\n- 3-8 sentence passage from published-quality source (journal, novel, essay) with specific names/dates/findings\n- Mix subtypes: main idea (2-3), author's purpose (1-2), inference (2-3), tone (0-1), structure (1-2)\n- Topics: natural science, social science, literature, history, arts/humanities (use each)\n- Distractors: surface-level misreading, wrong scope, opposite interpretation\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"text_comprehension\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[data_analysis]\n- Passage describing realistic data/study results requiring INTERPRETATION, not lookup\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Data patterns: diminishing returns, optimal values, conditional relationships\n- Distractors: wrong column, wrong trend direction, correlation vs causation confusion\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"data_analysis\",\"passage\":\"...\",\"visual\":{...},\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[claim_support]\n- Mix: \"Which finding supports/undermines the claim?\" AND \"Which quotation illustrates the claim?\"\n- Claims must be specific and testable. Distractors are topically related but don't address the specific mechanism.\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"claim_support\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[standard_english_conventions]\n- Passage with [underlined: text here] portion. Student picks best replacement.\n- Cover: subject-verb agreement, pronouns, verb tense, parallel structure, modifiers, commas, semicolons, apostrophes, sentence boundaries\n- \"NO CHANGE\" (choice A) correct ~25% of the time, NOT always\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"standard_english_conventions\",\"passage\":\"...with [underlined: ...]...\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[transitions]\n- Passage with ______ blank connecting two ideas. All 4 choices grammatically correct; distinction is logical.\n- Cover: contrast, addition, cause/effect, example, sequence, concession\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"transitions\",\"passage\":\"...______ ...\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[rhetorical_synthesis]\n- Notes format: \"While researching [topic], a student has taken these notes:\\n- Note 1\\n- Note 2\\n- Note 3\"\n- Specific rhetorical goal stated. All 4 choices are grammatical sentences using the notes; only 1 achieves the goal.\n- Goals: emphasize difference, highlight finding, introduce topic, summarize conclusion, present cause-effect\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"rhetorical_synthesis\",\"passage\":\"...\",\"prompt\":\"The student wants to [goal]. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\nOutput ONLY a valid JSON array of EXACTLY 5 objects. No markdown. Count them before outputting.", "text": "[{\"id\":\"sat99-rw-m2-q22\",\"number\":22,\"type\":\"standard_english_conventions\",\"passage\":\"The archaeologist Kathleen Kenyon, whose excavations at Jericho in the 1950s revolutionized the field's understanding of Neolithic settlement, [underlined: are] now regarded as a pioneer of stratigraphic methodology.\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"is\",\"C\":\"were\",\"D\":\"being\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m2-q23\",\"number\":23,\"type\":\"standard_english_conventions\",\"passage\":\"Marine biologists studying the Great Barrier Reef have documented a surprising resilience in certain coral species; [underlined: however these] organisms continue to reproduce even after repeated bleaching events.\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"however, these\",\"C\":\"however: these\",\"D\":\"however these,\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m2-q24\",\"number\":24,\"type\":\"standard_english_conventions\",\"passage\":\"When composing her final symphony, Florence Price drew on spirituals, European classical forms, and [underlined: she incorporated syncopated rhythms] characteristic of African American folk traditions.\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"incorporating syncopated rhythms\",\"C\":\"syncopated rhythms\",\"D\":\"syncopated rhythms were incorporated\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m2-q25\",\"number\":25,\"type\":\"standard_english_conventions\",\"passage\":\"The committee's decision to fund the observatory surprised many researchers [underlined: who's] grant proposals had been rejected in earlier rounds of review.\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"whose\",\"C\":\"whos'\",\"D\":\"whom's\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m2-q26\",\"number\":26,\"type\":\"standard_english_conventions\",\"passage\":\"By the time the expedition reached the base camp on the southern glacier, the team [underlined: has endured] three weeks of subzero temperatures and persistent whiteout conditions.\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"endures\",\"C\":\"is enduring\",\"D\":\"had endured\"},\"answer_format\":\"mcq\",\"correct_answer\":\"D\"}]" }, { "costUsd": 0.082365, "model": "opus", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 4 SAT Reading & Writing questions.\nType: transitions | Test: 99 | Unit: 2 | Start Q#: 27\n\nMANDATORY: Your JSON array MUST contain exactly 4 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nFollow the instructions for your type:\n\n[vocabulary_in_context]\n- 2-3 sentence passage with ______ blank from specific academic domains (molecular biology, Renaissance art, behavioral economics, astrophysics, political philosophy)\n- 4 college-level word choices from the same semantic field\n- Correct word is most PRECISE for context, not most common synonym\n- ID format: sat{testNumber}-rw-m{mod}-q{N}\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"vocabulary_in_context\",\"passage\":\"...with ______...\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[text_comprehension]\n- 3-8 sentence passage from published-quality source (journal, novel, essay) with specific names/dates/findings\n- Mix subtypes: main idea (2-3), author's purpose (1-2), inference (2-3), tone (0-1), structure (1-2)\n- Topics: natural science, social science, literature, history, arts/humanities (use each)\n- Distractors: surface-level misreading, wrong scope, opposite interpretation\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"text_comprehension\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[data_analysis]\n- Passage describing realistic data/study results requiring INTERPRETATION, not lookup\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Data patterns: diminishing returns, optimal values, conditional relationships\n- Distractors: wrong column, wrong trend direction, correlation vs causation confusion\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"data_analysis\",\"passage\":\"...\",\"visual\":{...},\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[claim_support]\n- Mix: \"Which finding supports/undermines the claim?\" AND \"Which quotation illustrates the claim?\"\n- Claims must be specific and testable. Distractors are topically related but don't address the specific mechanism.\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"claim_support\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[standard_english_conventions]\n- Passage with [underlined: text here] portion. Student picks best replacement.\n- Cover: subject-verb agreement, pronouns, verb tense, parallel structure, modifiers, commas, semicolons, apostrophes, sentence boundaries\n- \"NO CHANGE\" (choice A) correct ~25% of the time, NOT always\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"standard_english_conventions\",\"passage\":\"...with [underlined: ...]...\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[transitions]\n- Passage with ______ blank connecting two ideas. All 4 choices grammatically correct; distinction is logical.\n- Cover: contrast, addition, cause/effect, example, sequence, concession\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"transitions\",\"passage\":\"...______ ...\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[rhetorical_synthesis]\n- Notes format: \"While researching [topic], a student has taken these notes:\\n- Note 1\\n- Note 2\\n- Note 3\"\n- Specific rhetorical goal stated. All 4 choices are grammatical sentences using the notes; only 1 achieves the goal.\n- Goals: emphasize difference, highlight finding, introduce topic, summarize conclusion, present cause-effect\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"rhetorical_synthesis\",\"passage\":\"...\",\"prompt\":\"The student wants to [goal]. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\nOutput ONLY a valid JSON array of EXACTLY 4 objects. No markdown. Count them before outputting.", "text": "[{\"id\":\"sat99-rw-m2-q27\",\"number\":27,\"type\":\"transitions\",\"passage\":\"The axolotl, a salamander native to lakes near Mexico City, retains its larval features—including external gills and a finned tail—throughout adulthood, a phenomenon known as neoteny. Most amphibians undergo metamorphosis to develop lungs and adapt to terrestrial life. ______ the axolotl remains fully aquatic, breeding and feeding in the same form it had as a juvenile.\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"By contrast,\",\"B\":\"Likewise,\",\"C\":\"As a result,\",\"D\":\"For instance,\"},\"answer_format\":\"mcq\",\"correct_answer\":\"A\"},{\"id\":\"sat99-rw-m2-q28\",\"number\":28,\"type\":\"transitions\",\"passage\":\"In her 2019 study of urban tree canopies, ecologist Mariana Velez found that neighborhoods with greater tree cover recorded summer temperatures up to 4°C lower than adjacent treeless blocks. The cooling effect persisted even during prolonged heat waves, when air-conditioning demand spiked elsewhere. ______ Velez concluded that municipal tree-planting initiatives could meaningfully reduce both energy consumption and heat-related illness in dense city centers.\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"Nevertheless,\",\"B\":\"On these grounds,\",\"C\":\"In addition,\",\"D\":\"Granted,\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m2-q29\",\"number\":29,\"type\":\"transitions\",\"passage\":\"Composer Florence Price submitted her Symphony in E minor to the 1932 Wanamaker Foundation competition and won first prize, prompting the Chicago Symphony Orchestra to perform the work the following year. For decades afterward, however, her scores were largely forgotten and her name absent from concert programs. ______ in 2009, a renovation crew discovered a trove of Price's manuscripts in her abandoned summer home, sparking renewed scholarly and performance interest.\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"Consequently,\",\"B\":\"Similarly,\",\"C\":\"Then,\",\"D\":\"Meanwhile,\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m2-q30\",\"number\":30,\"type\":\"transitions\",\"passage\":\"Behavioral economist Sendhil Mullainathan has argued that scarcity—whether of time, money, or food—captures the mind, narrowing attention to the immediate shortage and degrading performance on unrelated tasks. Critics initially dismissed the claim as overstated, suggesting that observed deficits stemmed from chronic stress rather than scarcity itself. ______ subsequent experiments in which participants were temporarily induced to feel scarce, and then released from that state, replicated the cognitive effects within a single laboratory session.\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"Moreover,\",\"B\":\"For example,\",\"C\":\"In short,\",\"D\":\"However,\"},\"answer_format\":\"mcq\",\"correct_answer\":\"D\"}]" }, { "costUsd": 0.12909, "model": "opus", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 3 SAT Reading & Writing questions.\nType: rhetorical_synthesis | Test: 99 | Unit: 2 | Start Q#: 31\n\nMANDATORY: Your JSON array MUST contain exactly 3 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nFollow the instructions for your type:\n\n[vocabulary_in_context]\n- 2-3 sentence passage with ______ blank from specific academic domains (molecular biology, Renaissance art, behavioral economics, astrophysics, political philosophy)\n- 4 college-level word choices from the same semantic field\n- Correct word is most PRECISE for context, not most common synonym\n- ID format: sat{testNumber}-rw-m{mod}-q{N}\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"vocabulary_in_context\",\"passage\":\"...with ______...\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[text_comprehension]\n- 3-8 sentence passage from published-quality source (journal, novel, essay) with specific names/dates/findings\n- Mix subtypes: main idea (2-3), author's purpose (1-2), inference (2-3), tone (0-1), structure (1-2)\n- Topics: natural science, social science, literature, history, arts/humanities (use each)\n- Distractors: surface-level misreading, wrong scope, opposite interpretation\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"text_comprehension\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[data_analysis]\n- Passage describing realistic data/study results requiring INTERPRETATION, not lookup\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Data patterns: diminishing returns, optimal values, conditional relationships\n- Distractors: wrong column, wrong trend direction, correlation vs causation confusion\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"data_analysis\",\"passage\":\"...\",\"visual\":{...},\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[claim_support]\n- Mix: \"Which finding supports/undermines the claim?\" AND \"Which quotation illustrates the claim?\"\n- Claims must be specific and testable. Distractors are topically related but don't address the specific mechanism.\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"claim_support\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[standard_english_conventions]\n- Passage with [underlined: text here] portion. Student picks best replacement.\n- Cover: subject-verb agreement, pronouns, verb tense, parallel structure, modifiers, commas, semicolons, apostrophes, sentence boundaries\n- \"NO CHANGE\" (choice A) correct ~25% of the time, NOT always\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"standard_english_conventions\",\"passage\":\"...with [underlined: ...]...\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[transitions]\n- Passage with ______ blank connecting two ideas. All 4 choices grammatically correct; distinction is logical.\n- Cover: contrast, addition, cause/effect, example, sequence, concession\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"transitions\",\"passage\":\"...______ ...\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[rhetorical_synthesis]\n- Notes format: \"While researching [topic], a student has taken these notes:\\n- Note 1\\n- Note 2\\n- Note 3\"\n- Specific rhetorical goal stated. All 4 choices are grammatical sentences using the notes; only 1 achieves the goal.\n- Goals: emphasize difference, highlight finding, introduce topic, summarize conclusion, present cause-effect\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"rhetorical_synthesis\",\"passage\":\"...\",\"prompt\":\"The student wants to [goal]. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\nOutput ONLY a valid JSON array of EXACTLY 3 objects. No markdown. Count them before outputting.", "text": "[{\"id\":\"sat99-rw-m2-q31\",\"number\":31,\"type\":\"rhetorical_synthesis\",\"passage\":\"While researching the foraging behavior of bumblebees, a student has taken the following notes:\\n- Bumblebees (genus $Bombus$) visit a wide variety of flowering plants.\\n- A 2019 study by Lars Chittka observed bumblebees in an enclosed arena with artificial flowers.\\n- Some flowers offered high-concentration sucrose solutions; others offered low-concentration solutions.\\n- Bumblebees in the study revisited high-concentration flowers significantly more often than low-concentration ones.\\n- Chittka concluded that bumblebees can distinguish between reward magnitudes and adjust foraging accordingly.\",\"prompt\":\"The student wants to present the main finding of Chittka's 2019 study. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"Bumblebees, members of the genus $Bombus$, are known to visit a wide variety of flowering plants in natural environments.\",\"B\":\"In a 2019 study, Lars Chittka observed bumblebees in an enclosed arena containing artificial flowers with varying sucrose concentrations.\",\"C\":\"In Chittka's 2019 study, bumblebees revisited artificial flowers offering high-concentration sucrose solutions significantly more often than those offering low-concentration solutions.\",\"D\":\"Chittka's research suggests that the foraging behavior of bumblebees is shaped by their ability to perceive features of their environment.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m2-q32\",\"number\":32,\"type\":\"rhetorical_synthesis\",\"passage\":\"While researching the Italian Renaissance painter Sofonisba Anguissola, a student has taken the following notes:\\n- Sofonisba Anguissola (c. 1532–1625) was a Northern Italian painter from Cremona.\\n- She trained under Bernardino Campi, an unusual arrangement for women of her era.\\n- Her 1555 painting $The$ $Chess$ $Game$ depicts her three sisters playing chess.\\n- $The$ $Chess$ $Game$ is considered an early example of a genre scene featuring identifiable individuals.\\n- Most genre scenes of the period depicted anonymous, generalized figures rather than specific people.\",\"prompt\":\"The student wants to emphasize a way in which Anguissola's $The$ $Chess$ $Game$ differed from typical genre scenes of its time. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"Sofonisba Anguissola, who lived from approximately 1532 to 1625, was a painter from the Northern Italian city of Cremona.\",\"B\":\"Whereas most genre scenes of the period depicted anonymous, generalized figures, Anguissola's 1555 painting $The$ $Chess$ $Game$ portrayed her three sisters as identifiable individuals.\",\"C\":\"Anguissola trained under the painter Bernardino Campi, an arrangement that was unusual for women during the Italian Renaissance.\",\"D\":\"Anguissola's $The$ $Chess$ $Game$, completed in 1555, depicts her three sisters engaged in a game of chess.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m2-q33\",\"number\":33,\"type\":\"rhetorical_synthesis\",\"passage\":\"While researching the formation of Saturn's rings, a student has taken the following notes:\\n- Saturn's rings are composed primarily of water ice particles ranging from micrometers to meters in size.\\n- One hypothesis proposes the rings formed roughly 4.5 billion years ago, alongside Saturn itself.\\n- A 2019 analysis of Cassini spacecraft data by Iess and colleagues estimated the rings' total mass.\\n- Based on the low mass and minimal contamination by interplanetary dust, Iess and colleagues concluded the rings are likely only 10 to 100 million years old.\\n- This estimate suggests the rings may have formed long after the planet itself, possibly from the breakup of an icy moon.\",\"prompt\":\"The student wants to introduce the topic of the research to an audience unfamiliar with Saturn's rings. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"Saturn's rings, which are composed primarily of water ice particles varying widely in size, have been the subject of recent scientific investigation into their age and origin.\",\"B\":\"Iess and colleagues analyzed Cassini spacecraft data in 2019 to estimate the total mass of Saturn's rings.\",\"C\":\"The low mass and minimal dust contamination of Saturn's rings led Iess and colleagues to conclude that the rings are 10 to 100 million years old.\",\"D\":\"Saturn's rings may have formed from the breakup of an icy moon long after Saturn itself coalesced roughly 4.5 billion years ago.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"A\"}]" } ], "results": [ { "attempts": 1, "index": 0, "output": { "costUsd": 0.08664, "model": "opus", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 5 SAT Reading & Writing questions.\nType: vocabulary_in_context | Test: 99 | Unit: 1 | Start Q#: 1\n\nMANDATORY: Your JSON array MUST contain exactly 5 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nFollow the instructions for your type:\n\n[vocabulary_in_context]\n- 2-3 sentence passage with ______ blank from specific academic domains (molecular biology, Renaissance art, behavioral economics, astrophysics, political philosophy)\n- 4 college-level word choices from the same semantic field\n- Correct word is most PRECISE for context, not most common synonym\n- ID format: sat{testNumber}-rw-m{mod}-q{N}\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"vocabulary_in_context\",\"passage\":\"...with ______...\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[text_comprehension]\n- 3-8 sentence passage from published-quality source (journal, novel, essay) with specific names/dates/findings\n- Mix subtypes: main idea (2-3), author's purpose (1-2), inference (2-3), tone (0-1), structure (1-2)\n- Topics: natural science, social science, literature, history, arts/humanities (use each)\n- Distractors: surface-level misreading, wrong scope, opposite interpretation\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"text_comprehension\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[data_analysis]\n- Passage describing realistic data/study results requiring INTERPRETATION, not lookup\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Data patterns: diminishing returns, optimal values, conditional relationships\n- Distractors: wrong column, wrong trend direction, correlation vs causation confusion\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"data_analysis\",\"passage\":\"...\",\"visual\":{...},\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[claim_support]\n- Mix: \"Which finding supports/undermines the claim?\" AND \"Which quotation illustrates the claim?\"\n- Claims must be specific and testable. Distractors are topically related but don't address the specific mechanism.\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"claim_support\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[standard_english_conventions]\n- Passage with [underlined: text here] portion. Student picks best replacement.\n- Cover: subject-verb agreement, pronouns, verb tense, parallel structure, modifiers, commas, semicolons, apostrophes, sentence boundaries\n- \"NO CHANGE\" (choice A) correct ~25% of the time, NOT always\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"standard_english_conventions\",\"passage\":\"...with [underlined: ...]...\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[transitions]\n- Passage with ______ blank connecting two ideas. All 4 choices grammatically correct; distinction is logical.\n- Cover: contrast, addition, cause/effect, example, sequence, concession\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"transitions\",\"passage\":\"...______ ...\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[rhetorical_synthesis]\n- Notes format: \"While researching [topic], a student has taken these notes:\\n- Note 1\\n- Note 2\\n- Note 3\"\n- Specific rhetorical goal stated. All 4 choices are grammatical sentences using the notes; only 1 achieves the goal.\n- Goals: emphasize difference, highlight finding, introduce topic, summarize conclusion, present cause-effect\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"rhetorical_synthesis\",\"passage\":\"...\",\"prompt\":\"The student wants to [goal]. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\nOutput ONLY a valid JSON array of EXACTLY 5 objects. No markdown. Count them before outputting.", "text": "[{\"id\":\"sat99-rw-m1-q1\",\"number\":1,\"type\":\"vocabulary_in_context\",\"passage\":\"In his 1503 treatise on proportion, the Florentine sculptor argued that classical statuary achieved its effects not through rigid adherence to mathematical ratios but through subtle departures from them. A slight elongation of the torso, a marginal tilt of the head—these deviations, he claimed, ______ the figure with a vitality that strict symmetry could never produce.\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"imbued\",\"B\":\"contaminated\",\"C\":\"stocked\",\"D\":\"appointed\"},\"answer_format\":\"mcq\",\"correct_answer\":\"A\"},{\"id\":\"sat99-rw-m1-q2\",\"number\":2,\"type\":\"vocabulary_in_context\",\"passage\":\"Behavioral economist Sendhil Mullainathan has argued that scarcity does not merely shape decisions—it ______ cognition itself. When the poor must constantly calculate trade-offs that the wealthy never notice, mental bandwidth is consumed by the arithmetic of survival, leaving fewer resources for long-term planning.\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"designates\",\"B\":\"restructures\",\"C\":\"transcribes\",\"D\":\"deputizes\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m1-q3\",\"number\":3,\"type\":\"vocabulary_in_context\",\"passage\":\"The discovery of FOXP2, often misleadingly dubbed the ''language gene,'' has been widely ______ in popular science writing. Though mutations in the gene do disrupt speech production in humans, researchers caution that it is one component of a vast regulatory network rather than a singular switch governing linguistic capacity.\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"repudiated\",\"B\":\"curtailed\",\"C\":\"oversimplified\",\"D\":\"fabricated\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m1-q4\",\"number\":4,\"type\":\"vocabulary_in_context\",\"passage\":\"Because Population III stars—the hypothetical first generation formed from primordial hydrogen and helium—have never been directly observed, their properties must be ______ from indirect evidence such as the chemical signatures of ancient, metal-poor stars still visible in the Milky Way's halo.\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"extracted\",\"B\":\"inferred\",\"C\":\"fabricated\",\"D\":\"disclosed\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m1-q5\",\"number\":5,\"type\":\"vocabulary_in_context\",\"passage\":\"In $\\\\textit{On Liberty}$, John Stuart Mill worried that majority opinion, when unchecked, could exert a form of coercion more insidious than any statute. Public disapproval, he argued, could ______ dissent so thoroughly that heterodox ideas would wither before they could even be articulated.\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"stifle\",\"B\":\"rebuke\",\"C\":\"dispute\",\"D\":\"evade\"},\"answer_format\":\"mcq\",\"correct_answer\":\"A\"}]" }, "status": "ok" }, { "attempts": 1, "index": 1, "output": { "costUsd": 0.280965, "model": "opus", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 8 SAT Reading & Writing questions.\nType: text_comprehension | Test: 99 | Unit: 1 | Start Q#: 6\n\nMANDATORY: Your JSON array MUST contain exactly 8 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nFollow the instructions for your type:\n\n[vocabulary_in_context]\n- 2-3 sentence passage with ______ blank from specific academic domains (molecular biology, Renaissance art, behavioral economics, astrophysics, political philosophy)\n- 4 college-level word choices from the same semantic field\n- Correct word is most PRECISE for context, not most common synonym\n- ID format: sat{testNumber}-rw-m{mod}-q{N}\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"vocabulary_in_context\",\"passage\":\"...with ______...\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[text_comprehension]\n- 3-8 sentence passage from published-quality source (journal, novel, essay) with specific names/dates/findings\n- Mix subtypes: main idea (2-3), author's purpose (1-2), inference (2-3), tone (0-1), structure (1-2)\n- Topics: natural science, social science, literature, history, arts/humanities (use each)\n- Distractors: surface-level misreading, wrong scope, opposite interpretation\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"text_comprehension\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[data_analysis]\n- Passage describing realistic data/study results requiring INTERPRETATION, not lookup\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Data patterns: diminishing returns, optimal values, conditional relationships\n- Distractors: wrong column, wrong trend direction, correlation vs causation confusion\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"data_analysis\",\"passage\":\"...\",\"visual\":{...},\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[claim_support]\n- Mix: \"Which finding supports/undermines the claim?\" AND \"Which quotation illustrates the claim?\"\n- Claims must be specific and testable. Distractors are topically related but don't address the specific mechanism.\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"claim_support\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[standard_english_conventions]\n- Passage with [underlined: text here] portion. Student picks best replacement.\n- Cover: subject-verb agreement, pronouns, verb tense, parallel structure, modifiers, commas, semicolons, apostrophes, sentence boundaries\n- \"NO CHANGE\" (choice A) correct ~25% of the time, NOT always\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"standard_english_conventions\",\"passage\":\"...with [underlined: ...]...\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[transitions]\n- Passage with ______ blank connecting two ideas. All 4 choices grammatically correct; distinction is logical.\n- Cover: contrast, addition, cause/effect, example, sequence, concession\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"transitions\",\"passage\":\"...______ ...\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[rhetorical_synthesis]\n- Notes format: \"While researching [topic], a student has taken these notes:\\n- Note 1\\n- Note 2\\n- Note 3\"\n- Specific rhetorical goal stated. All 4 choices are grammatical sentences using the notes; only 1 achieves the goal.\n- Goals: emphasize difference, highlight finding, introduce topic, summarize conclusion, present cause-effect\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"rhetorical_synthesis\",\"passage\":\"...\",\"prompt\":\"The student wants to [goal]. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\nOutput ONLY a valid JSON array of EXACTLY 8 objects. No markdown. Count them before outputting.", "text": "[{\"id\":\"sat99-rw-m1-q6\",\"number\":6,\"type\":\"text_comprehension\",\"passage\":\"In her 2019 study of cooperative breeding in pied babblers (Turdoides bicolor), zoologist Amanda Ridley documented that subordinate adults who had previously received vocal encouragement from dominant pair members were significantly more likely to provision unrelated nestlings than those who had not. Ridley argues that this pattern complicates the standard kin-selection account, which holds that helping behaviors evolve primarily because they benefit close genetic relatives. The vocal exchanges, she suggests, function less as instructions than as a kind of social currency that helpers accumulate and later redeem through opportunities to feed and be associated with the brood.\",\"prompt\":\"Which choice best states the main idea of the text?\",\"choices\":{\"A\":\"Ridley's research demonstrates that kin selection plays no measurable role in the cooperative behavior of pied babblers.\",\"B\":\"Ridley's findings suggest that social interactions, rather than genetic relatedness alone, may shape helping behavior in pied babblers.\",\"C\":\"Ridley discovered that dominant pied babblers use vocal commands to direct unrelated subordinates toward specific nestlings.\",\"D\":\"Ridley concludes that subordinate pied babblers provision nestlings primarily to gain reproductive opportunities of their own.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m1-q7\",\"number\":7,\"type\":\"text_comprehension\",\"passage\":\"The following text is adapted from Edith Wharton's 1905 novel The House of Mirth. Lily Bart, a young woman of limited means, has just encountered an acquaintance at a train station.\\n\\nSelden paused in surprise. In the afternoon rush of the Grand Central Station his eyes had been refreshed by the sight of Miss Lily Bart. It was a Monday in early September, and he was returning to his work from a hurried dip into the country; but what was Miss Bart doing in town at that season? If she had appeared to be catching a train, he might have inferred that he had come on her in the act of transition between one and another of the country-houses which disputed her presence after the close of the Newport season; but her desultory air perplexed him. She stood apart from the crowd, letting it drift by her to the platform or the street, and wearing an air of irresolution which might, as he surmised, be the mask of a very definite purpose.\",\"prompt\":\"Which choice best describes the function of the final sentence in the text as a whole?\",\"choices\":{\"A\":\"It resolves Selden's earlier confusion by clarifying Lily's reasons for being at the station.\",\"B\":\"It contrasts Lily's outward behavior with the more chaotic behavior of the surrounding crowd.\",\"C\":\"It introduces Selden's suspicion that Lily's apparent indecision may conceal a deliberate intention.\",\"D\":\"It suggests that Lily is unaware of Selden's presence and is preoccupied with private worries.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m1-q8\",\"number\":8,\"type\":\"text_comprehension\",\"passage\":\"Art historian Patricia Fortini Brown has long maintained that fifteenth-century Venetian painters approached classical antiquity differently from their Florentine contemporaries. Where Florentine artists such as Mantegna treated ancient ruins and inscriptions as authoritative models to be reconstructed with archaeological precision, Venetian painters like Gentile Bellini incorporated antique fragments into compositions that subordinated historical accuracy to civic narrative. For Brown, the resulting images are not failures of scholarship but deliberate acts of appropriation: Venice, lacking a Roman past of its own, fashioned one by absorbing the relics of others into a visual rhetoric of imperial continuity.\",\"prompt\":\"As used in the text, the phrase \\\"deliberate acts of appropriation\\\" most nearly suggests that Venetian painters\",\"choices\":{\"A\":\"borrowed antique elements carelessly because they lacked access to reliable archaeological sources.\",\"B\":\"intentionally repurposed classical material to construct a flattering historical identity for their city.\",\"C\":\"competed with Florentine artists by producing more historically accurate reconstructions of Roman monuments.\",\"D\":\"rejected classical sources entirely in favor of subjects drawn from Venetian civic life.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m1-q9\",\"number\":9,\"type\":\"text_comprehension\",\"passage\":\"Recent measurements from the Parker Solar Probe have revealed that the solar wind is punctuated by sharp, S-shaped reversals in the magnetic field, which mission scientists have termed \\\"switchbacks.\\\" Although these structures had been glimpsed by earlier spacecraft, their abundance close to the Sun was not anticipated by prevailing models of coronal heating. Justin Kasper, the mission's principal investigator, cautions that it remains unclear whether switchbacks are generated near the solar surface by small-scale magnetic reconnection events or arise farther out through turbulent interactions in the expanding wind. Distinguishing between these possibilities, he notes, will require coordinated observations from probes operating at different distances from the Sun.\",\"prompt\":\"Based on the text, what can most reasonably be inferred about the origin of switchbacks?\",\"choices\":{\"A\":\"Researchers have determined that switchbacks form primarily through magnetic reconnection at the Sun's surface.\",\"B\":\"Researchers consider switchbacks too rare near the Sun to be explained by current coronal heating models.\",\"C\":\"Researchers do not yet have observations sufficient to decide between competing explanations of how switchbacks form.\",\"D\":\"Researchers believe switchbacks are an artifact of the Parker Solar Probe's instruments rather than a genuine solar phenomenon.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m1-q10\",\"number\":10,\"type\":\"text_comprehension\",\"passage\":\"In The Federalist No. 10, James Madison concedes that factions, which he defines as groups united by interests adverse to the rights of others or to the permanent interests of the community, are an unavoidable feature of free societies. To suppress factions, he argues, would require either destroying the liberty that nourishes them or imposing a uniformity of opinion that is impossible among reasoning beings. Madison therefore turns from the question of preventing factions to that of controlling their effects, contending that an extended republic, by multiplying the variety of interests it contains, makes the formation of an oppressive majority less probable than it would be in a small, homogeneous democracy.\",\"prompt\":\"Which choice best describes the overall structure of the text?\",\"choices\":{\"A\":\"It defines a political problem, dismisses two proposed solutions, and advances an alternative remedy.\",\"B\":\"It presents a historical example, identifies its weaknesses, and proposes a return to earlier practices.\",\"C\":\"It introduces a controversial claim, surveys objections to it, and concludes that the claim cannot be defended.\",\"D\":\"It compares two competing forms of government and recommends a compromise between them.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"A\"},{\"id\":\"sat99-rw-m1-q11\",\"number\":11,\"type\":\"text_comprehension\",\"passage\":\"The following text is adapted from a 1903 essay by W. E. B. Du Bois.\\n\\nBetween me and the other world there is ever an unasked question: unasked by some through feelings of delicacy; by others through the difficulty of rightly framing it. All, nevertheless, flutter round it. They approach me in a half-hesitant sort of way, eye me curiously or compassionately, and then, instead of saying directly, How does it feel to be a problem? they say, I know an excellent colored man in my town; or, I fought at Mechanicsville; or, Do not these Southern outrages make your blood boil? At these I smile, or am interested, or reduce the boiling to a simmer, as the occasion may require. To the real question, How does it feel to be a problem? I answer seldom a word.\",\"prompt\":\"Which choice best describes the speaker's tone in the text?\",\"choices\":{\"A\":\"Wry restraint mingled with quiet weariness\",\"B\":\"Defiant outrage barely contained by politeness\",\"C\":\"Cheerful detachment about a trivial misunderstanding\",\"D\":\"Hopeful conviction that misunderstandings will soon be resolved\"},\"answer_format\":\"mcq\",\"correct_answer\":\"A\"},{\"id\":\"sat99-rw-m1-q12\",\"number\":12,\"type\":\"text_comprehension\",\"passage\":\"Behavioral economist Sendhil Mullainathan and his collaborators have argued that the experience of scarcity, whether of money, time, or calories, alters cognition in ways that conventional economic models fail to capture. In a series of experiments, participants asked to manage tight budgets performed better on tasks directly related to those budgets but worse on unrelated tests of fluid intelligence and self-control. Mullainathan describes this trade-off as a kind of cognitive \\\"tax\\\" levied by scarcity itself, one that may help explain why interventions designed to alleviate poverty often falter when they demand additional planning, paperwork, or attention from the very people they aim to assist.\",\"prompt\":\"Why does the author most likely describe scarcity as a cognitive \\\"tax\\\"?\",\"choices\":{\"A\":\"To suggest that governments should compensate individuals financially for the mental costs of being poor\",\"B\":\"To emphasize that scarcity imposes mental costs that can undermine programs requiring sustained effort from beneficiaries\",\"C\":\"To argue that conventional economic models accurately predict how people respond to financial constraints\",\"D\":\"To imply that people experiencing scarcity perform worse on every cognitive task than wealthier people do\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m1-q13\",\"number\":13,\"type\":\"text_comprehension\",\"passage\":\"Marine biologist Roger Payne's 1971 recordings of humpback whale songs are often credited with galvanizing public support for the international moratorium on commercial whaling adopted a decade later. Yet the recordings' influence cannot be attributed to scientific argument alone. By releasing the songs as a commercial album rather than confining them to academic journals, Payne placed the voices of whales within the same cultural space as popular music, inviting listeners to perceive the animals as expressive beings rather than as units of harvestable biomass. The conservation movement that followed drew as much from this aesthetic reframing as from the demographic data that researchers were simultaneously compiling.\",\"prompt\":\"Which choice best states the main purpose of the text?\",\"choices\":{\"A\":\"To document the scientific methods Payne used to record and analyze humpback whale vocalizations\",\"B\":\"To argue that the moratorium on commercial whaling was adopted primarily because of advances in population data\",\"C\":\"To suggest that Payne's recordings shaped public attitudes toward whales in part by reframing them aesthetically rather than only scientifically\",\"D\":\"To criticize the conservation movement for relying on emotional appeals rather than on rigorous scientific evidence\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"}]" }, "status": "ok" }, { "attempts": 1, "index": 2, "output": { "costUsd": 0.15984, "model": "opus", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 4 SAT Reading & Writing questions.\nType: data_analysis | Test: 99 | Unit: 1 | Start Q#: 14\n\nMANDATORY: Your JSON array MUST contain exactly 4 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nFollow the instructions for your type:\n\n[vocabulary_in_context]\n- 2-3 sentence passage with ______ blank from specific academic domains (molecular biology, Renaissance art, behavioral economics, astrophysics, political philosophy)\n- 4 college-level word choices from the same semantic field\n- Correct word is most PRECISE for context, not most common synonym\n- ID format: sat{testNumber}-rw-m{mod}-q{N}\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"vocabulary_in_context\",\"passage\":\"...with ______...\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[text_comprehension]\n- 3-8 sentence passage from published-quality source (journal, novel, essay) with specific names/dates/findings\n- Mix subtypes: main idea (2-3), author's purpose (1-2), inference (2-3), tone (0-1), structure (1-2)\n- Topics: natural science, social science, literature, history, arts/humanities (use each)\n- Distractors: surface-level misreading, wrong scope, opposite interpretation\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"text_comprehension\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[data_analysis]\n- Passage describing realistic data/study results requiring INTERPRETATION, not lookup\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Data patterns: diminishing returns, optimal values, conditional relationships\n- Distractors: wrong column, wrong trend direction, correlation vs causation confusion\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"data_analysis\",\"passage\":\"...\",\"visual\":{...},\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[claim_support]\n- Mix: \"Which finding supports/undermines the claim?\" AND \"Which quotation illustrates the claim?\"\n- Claims must be specific and testable. Distractors are topically related but don't address the specific mechanism.\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"claim_support\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[standard_english_conventions]\n- Passage with [underlined: text here] portion. Student picks best replacement.\n- Cover: subject-verb agreement, pronouns, verb tense, parallel structure, modifiers, commas, semicolons, apostrophes, sentence boundaries\n- \"NO CHANGE\" (choice A) correct ~25% of the time, NOT always\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"standard_english_conventions\",\"passage\":\"...with [underlined: ...]...\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[transitions]\n- Passage with ______ blank connecting two ideas. All 4 choices grammatically correct; distinction is logical.\n- Cover: contrast, addition, cause/effect, example, sequence, concession\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"transitions\",\"passage\":\"...______ ...\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[rhetorical_synthesis]\n- Notes format: \"While researching [topic], a student has taken these notes:\\n- Note 1\\n- Note 2\\n- Note 3\"\n- Specific rhetorical goal stated. All 4 choices are grammatical sentences using the notes; only 1 achieves the goal.\n- Goals: emphasize difference, highlight finding, introduce topic, summarize conclusion, present cause-effect\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"rhetorical_synthesis\",\"passage\":\"...\",\"prompt\":\"The student wants to [goal]. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\nOutput ONLY a valid JSON array of EXACTLY 4 objects. No markdown. Count them before outputting.", "text": "[{\"id\":\"sat99-rw-m1-q14\",\"number\":14,\"type\":\"data_analysis\",\"passage\":\"Soil ecologist Maren Holst investigated how earthworm density affects the decomposition rate of leaf litter in temperate deciduous forests. She established experimental plots with varying numbers of earthworms per square meter and measured the percentage of leaf mass lost over a 90-day period. Holst hypothesized that decomposition would increase with earthworm density but that the rate of increase would diminish at higher densities because of resource competition among the worms.\",\"visual\":{\"type\":\"table\",\"title\":\"Leaf Litter Decomposition by Earthworm Density (90 Days)\",\"headers\":[\"Earthworms per m²\",\"Leaf Mass Lost (%)\",\"Increase from Previous Level (%)\"],\"rows\":[[\"0\",\"12\",\"—\"],[\"20\",\"28\",\"16\"],[\"40\",\"41\",\"13\"],[\"60\",\"49\",\"8\"],[\"80\",\"53\",\"4\"]]},\"prompt\":\"Which choice best describes data from the table that support Holst's hypothesis?\",\"choices\":{\"A\":\"Decomposition was highest at 80 earthworms per m², and the increase from the previous density level was the smallest observed.\",\"B\":\"Decomposition increased steadily from 0 to 80 earthworms per m², and the increase from each previous level was identical across densities.\",\"C\":\"Decomposition was lowest at 0 earthworms per m², and the increase from 0 to 20 earthworms per m² was smaller than the increase from 60 to 80.\",\"D\":\"Decomposition peaked at 40 earthworms per m², after which leaf mass lost declined as density increased further.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"A\"},{\"id\":\"sat99-rw-m1-q15\",\"number\":15,\"type\":\"data_analysis\",\"passage\":\"A team of behavioral economists led by Rajiv Chandra examined whether the timing of a charitable appeal influences donation amounts. Participants were randomly assigned to receive a solicitation email either on the morning of payday, three days after payday, or ten days after payday. The researchers recorded the mean donation in each condition as well as the percentage of recipients who donated at all.\",\"visual\":{\"type\":\"table\",\"title\":\"Donation Behavior by Email Timing\",\"headers\":[\"Timing of Appeal\",\"Donation Rate (%)\",\"Mean Donation ($)\"],\"rows\":[[\"Morning of payday\",\"34\",\"22.50\"],[\"3 days after payday\",\"29\",\"18.75\"],[\"10 days after payday\",\"18\",\"11.20\"]]},\"prompt\":\"Which choice most accurately describes the relationship between appeal timing and donor behavior shown in the table?\",\"choices\":{\"A\":\"As the delay after payday increased, the donation rate rose while the mean donation amount fell.\",\"B\":\"As the delay after payday increased, both the donation rate and the mean donation amount declined.\",\"C\":\"The donation rate was highest ten days after payday, but the mean donation was highest on the morning of payday.\",\"D\":\"Appeals sent three days after payday produced both the highest donation rate and the highest mean donation.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m1-q16\",\"number\":16,\"type\":\"data_analysis\",\"passage\":\"Materials scientist Inés Beltrán tested the tensile strength of a novel bio-based polymer at several curing temperatures to determine an optimal processing condition. She noted that insufficient curing leaves polymer chains weakly bonded, whereas excessive heat can degrade the chains and reduce strength. Beltrán predicted that tensile strength would rise with temperature up to a threshold and then fall beyond it.\",\"visual\":{\"type\":\"table\",\"title\":\"Tensile Strength of Bio-Based Polymer by Curing Temperature\",\"headers\":[\"Curing Temperature (°C)\",\"Tensile Strength (MPa)\"],\"rows\":[[\"100\",\"42\"],[\"120\",\"58\"],[\"140\",\"71\"],[\"160\",\"65\"],[\"180\",\"49\"]]},\"prompt\":\"Which choice best describes data from the table that support Beltrán's prediction?\",\"choices\":{\"A\":\"Tensile strength rose continuously from 100 °C to 180 °C, confirming that higher curing temperatures always improve the polymer.\",\"B\":\"Tensile strength was lowest at 100 °C and highest at 180 °C, indicating that thermal degradation did not occur within the tested range.\",\"C\":\"Tensile strength increased from 100 °C to 140 °C and then decreased at 160 °C and 180 °C, indicating a peak within the tested range.\",\"D\":\"Tensile strength remained constant between 120 °C and 160 °C, suggesting that curing temperature has little effect in that interval.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m1-q17\",\"number\":17,\"type\":\"data_analysis\",\"passage\":\"In a study of cognitive load during multitasking, psychologist Theo Abedi asked participants to complete a primary reading comprehension task while performing a secondary task of varying difficulty. He measured reading accuracy (percentage of comprehension questions answered correctly) and reaction time on the secondary task. Abedi wanted to determine whether a moderately demanding secondary task would disrupt primary-task performance less than either a very easy or a very difficult one.\",\"visual\":{\"type\":\"table\",\"title\":\"Performance by Secondary-Task Difficulty\",\"headers\":[\"Secondary-Task Difficulty\",\"Reading Accuracy (%)\",\"Secondary-Task Reaction Time (ms)\"],\"rows\":[[\"Very easy\",\"74\",\"310\"],[\"Moderate\",\"88\",\"425\"],[\"Difficult\",\"61\",\"690\"]]},\"prompt\":\"Which choice best describes data from the table that are relevant to Abedi's research question?\",\"choices\":{\"A\":\"Reading accuracy was lower in the difficult condition than in the very easy condition, showing that harder secondary tasks always impair primary-task performance.\",\"B\":\"Reading accuracy was higher in the moderate condition than in either the very easy or the difficult condition, consistent with the idea that a moderately demanding secondary task disrupts reading least.\",\"C\":\"Reaction time on the secondary task was shortest in the very easy condition, indicating that participants focused most effectively on reading when the secondary task was simplest.\",\"D\":\"Reading accuracy and secondary-task reaction time both increased as difficulty rose, suggesting that the two measures track each other across conditions.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"}]" }, "status": "ok" }, { "attempts": 1, "index": 3, "output": { "costUsd": 0.150315, "model": "opus", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 4 SAT Reading & Writing questions.\nType: claim_support | Test: 99 | Unit: 1 | Start Q#: 18\n\nMANDATORY: Your JSON array MUST contain exactly 4 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nFollow the instructions for your type:\n\n[vocabulary_in_context]\n- 2-3 sentence passage with ______ blank from specific academic domains (molecular biology, Renaissance art, behavioral economics, astrophysics, political philosophy)\n- 4 college-level word choices from the same semantic field\n- Correct word is most PRECISE for context, not most common synonym\n- ID format: sat{testNumber}-rw-m{mod}-q{N}\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"vocabulary_in_context\",\"passage\":\"...with ______...\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[text_comprehension]\n- 3-8 sentence passage from published-quality source (journal, novel, essay) with specific names/dates/findings\n- Mix subtypes: main idea (2-3), author's purpose (1-2), inference (2-3), tone (0-1), structure (1-2)\n- Topics: natural science, social science, literature, history, arts/humanities (use each)\n- Distractors: surface-level misreading, wrong scope, opposite interpretation\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"text_comprehension\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[data_analysis]\n- Passage describing realistic data/study results requiring INTERPRETATION, not lookup\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Data patterns: diminishing returns, optimal values, conditional relationships\n- Distractors: wrong column, wrong trend direction, correlation vs causation confusion\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"data_analysis\",\"passage\":\"...\",\"visual\":{...},\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[claim_support]\n- Mix: \"Which finding supports/undermines the claim?\" AND \"Which quotation illustrates the claim?\"\n- Claims must be specific and testable. Distractors are topically related but don't address the specific mechanism.\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"claim_support\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[standard_english_conventions]\n- Passage with [underlined: text here] portion. Student picks best replacement.\n- Cover: subject-verb agreement, pronouns, verb tense, parallel structure, modifiers, commas, semicolons, apostrophes, sentence boundaries\n- \"NO CHANGE\" (choice A) correct ~25% of the time, NOT always\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"standard_english_conventions\",\"passage\":\"...with [underlined: ...]...\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[transitions]\n- Passage with ______ blank connecting two ideas. All 4 choices grammatically correct; distinction is logical.\n- Cover: contrast, addition, cause/effect, example, sequence, concession\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"transitions\",\"passage\":\"...______ ...\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[rhetorical_synthesis]\n- Notes format: \"While researching [topic], a student has taken these notes:\\n- Note 1\\n- Note 2\\n- Note 3\"\n- Specific rhetorical goal stated. All 4 choices are grammatical sentences using the notes; only 1 achieves the goal.\n- Goals: emphasize difference, highlight finding, introduce topic, summarize conclusion, present cause-effect\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"rhetorical_synthesis\",\"passage\":\"...\",\"prompt\":\"The student wants to [goal]. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\nOutput ONLY a valid JSON array of EXACTLY 4 objects. No markdown. Count them before outputting.", "text": "[{\"id\":\"sat99-rw-m1-q18\",\"number\":18,\"type\":\"claim_support\",\"passage\":\"A researcher studying the cognitive development of corvids (a family of birds including crows and ravens) has proposed that New Caledonian crows possess a form of causal reasoning that extends beyond simple trial-and-error learning. Specifically, the researcher claims that these birds can mentally simulate the physical consequences of an action before executing it, allowing them to solve novel tool-use problems on the first attempt without prior exposure.\",\"prompt\":\"Which finding, if true, would most directly support the researcher's claim?\",\"choices\":{\"A\":\"In laboratory settings, New Caledonian crows that observed other crows retrieving food with hooked sticks were subsequently able to use similar tools themselves after several practice sessions.\",\"B\":\"When presented with a novel apparatus requiring them to drop stones into a water-filled tube to raise a floating reward, naive New Caledonian crows selected stones of appropriate size and density on their first attempt, ignoring objects that would float.\",\"C\":\"New Caledonian crows in the wild have been documented manufacturing hooked tools from pandanus leaves, a behavior that varies regionally and appears to be transmitted across generations.\",\"D\":\"Neuroanatomical studies reveal that New Caledonian crows possess a higher density of neurons in the pallium than other corvid species of comparable body size.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m1-q19\",\"number\":19,\"type\":\"claim_support\",\"passage\":\"In her 2021 study of urban microclimates, geographer Lena Abernathy argues that the cooling effect of small neighborhood parks on surrounding city blocks is disproportionately large relative to their area—a phenomenon she attributes not to the parks' vegetation alone but to the way they disrupt the continuous surface of heat-absorbing pavement and buildings that characterize dense urban cores.\",\"prompt\":\"Which finding from Abernathy's study would most directly support her specific explanation for the cooling effect?\",\"choices\":{\"A\":\"Neighborhood parks measuring less than one hectare were associated with average temperature reductions of 1.8°C in adjacent blocks during summer afternoons.\",\"B\":\"Tree-canopy coverage within the parks correlated positively with the magnitude of temperature reduction observed in surrounding areas.\",\"C\":\"Paved plazas of equivalent size to the parks, when inserted into dense urban blocks, produced measurable cooling effects roughly 60 percent as large as those of vegetated parks.\",\"D\":\"Residents living within 200 meters of a neighborhood park reported lower rates of heat-related illness than residents living farther away during extreme heat events.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m1-q20\",\"number\":20,\"type\":\"claim_support\",\"passage\":\"In his critical study of the Victorian novel, literary scholar Marcus Holbrook contends that Elizabeth Gaskell's \\\"North and South\\\" (1855) departs from the conventions of the industrial novel by granting its female protagonist, Margaret Hale, a role as economic arbiter rather than as a passive moral influence. Holbrook argues that Margaret's decisions shape the material fortunes of the novel's central mill, a reversal of the typical Victorian pattern in which women inspire male action without directly determining business outcomes.\",\"prompt\":\"Which quotation from \\\"North and South\\\" would most effectively illustrate Holbrook's claim?\",\"choices\":{\"A\":\"\\\"Margaret felt, as she watched the crowds streaming from the mill gates, a sorrow she could scarcely name, a pity for the weary faces that seemed to reproach her own ease.\\\"\",\"B\":\"\\\"Mr. Thornton's manner toward her had grown gentler since their last conversation, and she was aware that her words had, in some measure, softened his views on the conduct of his workmen.\\\"\",\"C\":\"\\\"It was Margaret's capital, inherited from Mr. Bell and offered upon terms of her own devising, that permitted Thornton to reopen Marlborough Mills when no bank in Milton would advance him a shilling.\\\"\",\"D\":\"\\\"She thought often of the North, of its smoke and clangor, and wondered whether the gentleness of the South had not unfitted her for a world in which men like Thornton labored and contended.\\\"\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m1-q21\",\"number\":21,\"type\":\"claim_support\",\"passage\":\"Archaeologist Priya Chandran has argued that the residents of Çatalhöyük, a Neolithic settlement in southern Anatolia occupied from roughly 7100 to 5950 BCE, maintained a remarkably egalitarian social structure throughout the settlement's occupation, with no single household accumulating significantly greater wealth or status than its neighbors. Critics have suggested that this apparent egalitarianism may reflect the limitations of the archaeological record rather than the actual social organization of the community.\",\"prompt\":\"Which finding, if true, would most directly undermine Chandran's claim?\",\"choices\":{\"A\":\"Isotopic analysis of human remains from Çatalhöyük indicates that dietary composition varied little across individuals, suggesting broadly similar access to food resources.\",\"B\":\"Excavations at a contemporaneous settlement 200 kilometers to the east have revealed clear evidence of status differentiation, including elite burials with imported obsidian goods.\",\"C\":\"Detailed analysis of the distribution of rare imported materials across Çatalhöyük households reveals that a small cluster of dwellings consistently contained ten times the quantity found elsewhere in the settlement.\",\"D\":\"The layout of Çatalhöyük, with its densely packed houses sharing walls and accessed through roof openings, has been interpreted as evidence of strong communal identity.\",\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"}]" }, "status": "ok" }, { "attempts": 1, "index": 4, "output": { "costUsd": 0.08124, "model": "opus", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 5 SAT Reading & Writing questions.\nType: standard_english_conventions | Test: 99 | Unit: 1 | Start Q#: 22\n\nMANDATORY: Your JSON array MUST contain exactly 5 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nFollow the instructions for your type:\n\n[vocabulary_in_context]\n- 2-3 sentence passage with ______ blank from specific academic domains (molecular biology, Renaissance art, behavioral economics, astrophysics, political philosophy)\n- 4 college-level word choices from the same semantic field\n- Correct word is most PRECISE for context, not most common synonym\n- ID format: sat{testNumber}-rw-m{mod}-q{N}\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"vocabulary_in_context\",\"passage\":\"...with ______...\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[text_comprehension]\n- 3-8 sentence passage from published-quality source (journal, novel, essay) with specific names/dates/findings\n- Mix subtypes: main idea (2-3), author's purpose (1-2), inference (2-3), tone (0-1), structure (1-2)\n- Topics: natural science, social science, literature, history, arts/humanities (use each)\n- Distractors: surface-level misreading, wrong scope, opposite interpretation\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"text_comprehension\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[data_analysis]\n- Passage describing realistic data/study results requiring INTERPRETATION, not lookup\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Data patterns: diminishing returns, optimal values, conditional relationships\n- Distractors: wrong column, wrong trend direction, correlation vs causation confusion\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"data_analysis\",\"passage\":\"...\",\"visual\":{...},\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[claim_support]\n- Mix: \"Which finding supports/undermines the claim?\" AND \"Which quotation illustrates the claim?\"\n- Claims must be specific and testable. Distractors are topically related but don't address the specific mechanism.\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"claim_support\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[standard_english_conventions]\n- Passage with [underlined: text here] portion. Student picks best replacement.\n- Cover: subject-verb agreement, pronouns, verb tense, parallel structure, modifiers, commas, semicolons, apostrophes, sentence boundaries\n- \"NO CHANGE\" (choice A) correct ~25% of the time, NOT always\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"standard_english_conventions\",\"passage\":\"...with [underlined: ...]...\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[transitions]\n- Passage with ______ blank connecting two ideas. All 4 choices grammatically correct; distinction is logical.\n- Cover: contrast, addition, cause/effect, example, sequence, concession\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"transitions\",\"passage\":\"...______ ...\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[rhetorical_synthesis]\n- Notes format: \"While researching [topic], a student has taken these notes:\\n- Note 1\\n- Note 2\\n- Note 3\"\n- Specific rhetorical goal stated. All 4 choices are grammatical sentences using the notes; only 1 achieves the goal.\n- Goals: emphasize difference, highlight finding, introduce topic, summarize conclusion, present cause-effect\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"rhetorical_synthesis\",\"passage\":\"...\",\"prompt\":\"The student wants to [goal]. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\nOutput ONLY a valid JSON array of EXACTLY 5 objects. No markdown. Count them before outputting.", "text": "[{\"id\":\"sat99-rw-m1-q22\",\"number\":22,\"type\":\"standard_english_conventions\",\"passage\":\"The sculptor Louise Nevelson, who emigrated from what is now Ukraine to the United States in 1905, became renowned for her monochromatic wooden assemblages; these works, often painted entirely black or white, [underlined: transforms] discarded objects into monumental compositions.\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"transform\",\"C\":\"is transforming\",\"D\":\"transforming\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m1-q23\",\"number\":23,\"type\":\"standard_english_conventions\",\"passage\":\"Marine biologist Sylvia Earle has spent decades advocating for ocean conservation. Her research expeditions have taken her to depths few humans have reached, and [underlined: they're findings have] reshaped public understanding of deep-sea ecosystems.\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"their findings have\",\"C\":\"there findings have\",\"D\":\"they're findings has\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m1-q24\",\"number\":24,\"type\":\"standard_english_conventions\",\"passage\":\"In traditional Japanese $kintsugi$, artisans repair broken pottery with lacquer mixed with powdered gold, silver, or [underlined: platinum, the technique treats] breakage and repair as part of an object's history rather than something to disguise.\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"platinum the technique treats\",\"C\":\"platinum; the technique treats\",\"D\":\"platinum, treating\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m1-q25\",\"number\":25,\"type\":\"standard_english_conventions\",\"passage\":\"Economist Esther Duflo, who shared the 2019 Nobel Prize for her experimental approach to alleviating global poverty, argues that small-scale randomized trials can reveal which interventions actually [underlined: works] in specific local contexts.\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"is working\",\"C\":\"has worked\",\"D\":\"work\"},\"answer_format\":\"mcq\",\"correct_answer\":\"D\"},{\"id\":\"sat99-rw-m1-q26\",\"number\":26,\"type\":\"standard_english_conventions\",\"passage\":\"By the time astronomers confirmed the existence of the first exoplanet orbiting a sun-like star in 1995, decades of speculation about worlds beyond our solar [underlined: system, had] given way to a concrete, observable field of research.\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"system had\",\"C\":\"system; had\",\"D\":\"system: had\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"}]" }, "status": "ok" }, { "attempts": 1, "index": 5, "output": { "costUsd": 0.07734, "model": "opus", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 4 SAT Reading & Writing questions.\nType: transitions | Test: 99 | Unit: 1 | Start Q#: 27\n\nMANDATORY: Your JSON array MUST contain exactly 4 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nFollow the instructions for your type:\n\n[vocabulary_in_context]\n- 2-3 sentence passage with ______ blank from specific academic domains (molecular biology, Renaissance art, behavioral economics, astrophysics, political philosophy)\n- 4 college-level word choices from the same semantic field\n- Correct word is most PRECISE for context, not most common synonym\n- ID format: sat{testNumber}-rw-m{mod}-q{N}\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"vocabulary_in_context\",\"passage\":\"...with ______...\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[text_comprehension]\n- 3-8 sentence passage from published-quality source (journal, novel, essay) with specific names/dates/findings\n- Mix subtypes: main idea (2-3), author's purpose (1-2), inference (2-3), tone (0-1), structure (1-2)\n- Topics: natural science, social science, literature, history, arts/humanities (use each)\n- Distractors: surface-level misreading, wrong scope, opposite interpretation\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"text_comprehension\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[data_analysis]\n- Passage describing realistic data/study results requiring INTERPRETATION, not lookup\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Data patterns: diminishing returns, optimal values, conditional relationships\n- Distractors: wrong column, wrong trend direction, correlation vs causation confusion\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"data_analysis\",\"passage\":\"...\",\"visual\":{...},\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[claim_support]\n- Mix: \"Which finding supports/undermines the claim?\" AND \"Which quotation illustrates the claim?\"\n- Claims must be specific and testable. Distractors are topically related but don't address the specific mechanism.\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"claim_support\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[standard_english_conventions]\n- Passage with [underlined: text here] portion. Student picks best replacement.\n- Cover: subject-verb agreement, pronouns, verb tense, parallel structure, modifiers, commas, semicolons, apostrophes, sentence boundaries\n- \"NO CHANGE\" (choice A) correct ~25% of the time, NOT always\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"standard_english_conventions\",\"passage\":\"...with [underlined: ...]...\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[transitions]\n- Passage with ______ blank connecting two ideas. All 4 choices grammatically correct; distinction is logical.\n- Cover: contrast, addition, cause/effect, example, sequence, concession\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"transitions\",\"passage\":\"...______ ...\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[rhetorical_synthesis]\n- Notes format: \"While researching [topic], a student has taken these notes:\\n- Note 1\\n- Note 2\\n- Note 3\"\n- Specific rhetorical goal stated. All 4 choices are grammatical sentences using the notes; only 1 achieves the goal.\n- Goals: emphasize difference, highlight finding, introduce topic, summarize conclusion, present cause-effect\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"rhetorical_synthesis\",\"passage\":\"...\",\"prompt\":\"The student wants to [goal]. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\nOutput ONLY a valid JSON array of EXACTLY 4 objects. No markdown. Count them before outputting.", "text": "[{\"id\":\"sat99-rw-m1-q27\",\"number\":27,\"type\":\"transitions\",\"passage\":\"The axolotl, a salamander native to lakes near Mexico City, retains its larval features—including external gills and a finned tail—throughout its adult life, a phenomenon known as neoteny. Most salamanders undergo metamorphosis, shedding these juvenile traits as they mature. The axolotl, ______ remains perpetually youthful in appearance, a trait that has made it a valuable subject for studies of regeneration and development.\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"by contrast,\",\"B\":\"for instance,\",\"C\":\"as a result,\",\"D\":\"in addition,\"},\"answer_format\":\"mcq\",\"correct_answer\":\"A\"},{\"id\":\"sat99-rw-m1-q28\",\"number\":28,\"type\":\"transitions\",\"passage\":\"Archaeologist Sarah Parcak uses satellite imagery to identify buried archaeological sites, detecting subtle variations in soil and vegetation that hint at structures beneath the surface. Traditional excavation methods can take years and require substantial funding to survey even a small region. Parcak's technique, ______ allows researchers to map hundreds of potential sites in a fraction of the time, dramatically accelerating the pace of discovery.\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"nevertheless,\",\"B\":\"likewise,\",\"C\":\"however,\",\"D\":\"for example,\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m1-q29\",\"number\":29,\"type\":\"transitions\",\"passage\":\"Economist Esther Duflo and her colleagues tested whether small financial incentives could increase childhood immunization rates in rural Rajasthan. Villages offering a kilogram of lentils per vaccination saw immunization rates rise from 6 percent to 38 percent. ______ the program proved more cost-effective than conventional outreach, because the higher participation reduced the per-child expense of operating each mobile clinic.\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"Nonetheless,\",\"B\":\"Similarly,\",\"C\":\"Moreover,\",\"D\":\"Instead,\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m1-q30\",\"number\":30,\"type\":\"transitions\",\"passage\":\"In the late 1870s, painter Mary Cassatt was invited by Edgar Degas to exhibit with the Impressionists in Paris, a group then regarded by critics as radical and unpolished. Cassatt's decision to align herself with these artists cost her the patronage of several conservative American collectors. ______ the affiliation gave her access to a circle of experimental printmakers whose techniques would reshape her own approach to composition.\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"Likewise,\",\"B\":\"In turn,\",\"C\":\"Therefore,\",\"D\":\"Still,\"},\"answer_format\":\"mcq\",\"correct_answer\":\"D\"}]" }, "status": "ok" }, { "attempts": 1, "index": 6, "output": { "costUsd": 0.121365, "model": "opus", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 3 SAT Reading & Writing questions.\nType: rhetorical_synthesis | Test: 99 | Unit: 1 | Start Q#: 31\n\nMANDATORY: Your JSON array MUST contain exactly 3 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nFollow the instructions for your type:\n\n[vocabulary_in_context]\n- 2-3 sentence passage with ______ blank from specific academic domains (molecular biology, Renaissance art, behavioral economics, astrophysics, political philosophy)\n- 4 college-level word choices from the same semantic field\n- Correct word is most PRECISE for context, not most common synonym\n- ID format: sat{testNumber}-rw-m{mod}-q{N}\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"vocabulary_in_context\",\"passage\":\"...with ______...\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[text_comprehension]\n- 3-8 sentence passage from published-quality source (journal, novel, essay) with specific names/dates/findings\n- Mix subtypes: main idea (2-3), author's purpose (1-2), inference (2-3), tone (0-1), structure (1-2)\n- Topics: natural science, social science, literature, history, arts/humanities (use each)\n- Distractors: surface-level misreading, wrong scope, opposite interpretation\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"text_comprehension\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[data_analysis]\n- Passage describing realistic data/study results requiring INTERPRETATION, not lookup\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Data patterns: diminishing returns, optimal values, conditional relationships\n- Distractors: wrong column, wrong trend direction, correlation vs causation confusion\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"data_analysis\",\"passage\":\"...\",\"visual\":{...},\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[claim_support]\n- Mix: \"Which finding supports/undermines the claim?\" AND \"Which quotation illustrates the claim?\"\n- Claims must be specific and testable. Distractors are topically related but don't address the specific mechanism.\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"claim_support\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[standard_english_conventions]\n- Passage with [underlined: text here] portion. Student picks best replacement.\n- Cover: subject-verb agreement, pronouns, verb tense, parallel structure, modifiers, commas, semicolons, apostrophes, sentence boundaries\n- \"NO CHANGE\" (choice A) correct ~25% of the time, NOT always\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"standard_english_conventions\",\"passage\":\"...with [underlined: ...]...\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[transitions]\n- Passage with ______ blank connecting two ideas. All 4 choices grammatically correct; distinction is logical.\n- Cover: contrast, addition, cause/effect, example, sequence, concession\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"transitions\",\"passage\":\"...______ ...\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[rhetorical_synthesis]\n- Notes format: \"While researching [topic], a student has taken these notes:\\n- Note 1\\n- Note 2\\n- Note 3\"\n- Specific rhetorical goal stated. All 4 choices are grammatical sentences using the notes; only 1 achieves the goal.\n- Goals: emphasize difference, highlight finding, introduce topic, summarize conclusion, present cause-effect\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"rhetorical_synthesis\",\"passage\":\"...\",\"prompt\":\"The student wants to [goal]. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\nOutput ONLY a valid JSON array of EXACTLY 3 objects. No markdown. Count them before outputting.", "text": "[{\"id\":\"sat99-rw-m1-q31\",\"number\":31,\"type\":\"rhetorical_synthesis\",\"passage\":\"While researching the migratory patterns of the Arctic tern, a student has taken these notes:\\n- The Arctic tern (Sterna paradisaea) is a small seabird with a wingspan of about 75 cm.\\n- It breeds in Arctic regions during the Northern Hemisphere summer.\\n- It winters in Antarctic waters during the Southern Hemisphere summer.\\n- A 2010 tracking study by Egevang et al. found terns travel roughly 71,000 km annually.\\n- This is the longest known migration of any animal on Earth.\",\"prompt\":\"The student wants to emphasize the remarkable nature of the Arctic tern's migration distance. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"The Arctic tern, a small seabird with a 75 cm wingspan, breeds in the Arctic and winters in Antarctic waters.\",\"B\":\"A 2010 tracking study by Egevang et al. revealed that the Arctic tern travels roughly 71,000 km annually—the longest known migration of any animal on Earth.\",\"C\":\"The Arctic tern breeds in Arctic regions during the Northern Hemisphere summer and migrates to Antarctic waters for the Southern Hemisphere summer.\",\"D\":\"Egevang and colleagues conducted a tracking study of the Arctic tern, a seabird whose scientific name is Sterna paradisaea, in 2010.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m1-q32\",\"number\":32,\"type\":\"rhetorical_synthesis\",\"passage\":\"While researching the painter Artemisia Gentileschi, a student has taken these notes:\\n- Artemisia Gentileschi (1593–1656) was an Italian Baroque painter.\\n- She trained in her father Orazio's workshop in Rome.\\n- In 1616, she became the first woman admitted to the Accademia delle Arti del Disegno in Florence.\\n- Her painting Judith Slaying Holofernes (c. 1620) is held by the Uffizi Gallery.\\n- Art historians cite the work for its dramatic chiaroscuro and unflinching depiction of violence.\",\"prompt\":\"The student wants to introduce Artemisia Gentileschi to an audience unfamiliar with her work. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"Judith Slaying Holofernes, painted around 1620, is noted by art historians for its dramatic chiaroscuro and unflinching depiction of violence.\",\"B\":\"In 1616, Artemisia Gentileschi became the first woman admitted to the Accademia delle Arti del Disegno in Florence.\",\"C\":\"Artemisia Gentileschi (1593–1656) was an Italian Baroque painter whose works, such as Judith Slaying Holofernes, are celebrated for their dramatic intensity.\",\"D\":\"Trained in her father Orazio's Roman workshop, Artemisia Gentileschi later moved to Florence, where the Uffizi Gallery now holds one of her paintings.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m1-q33\",\"number\":33,\"type\":\"rhetorical_synthesis\",\"passage\":\"While researching a behavioral economics experiment, a student has taken these notes:\\n- Researchers Dan Ariely and colleagues studied how anchoring affects willingness to pay.\\n- Participants were asked to write down the last two digits of their Social Security number.\\n- They were then asked to bid on items such as wine, chocolates, and electronics.\\n- Participants with higher digit pairs (80–99) bid up to 346% more than those with lower pairs (00–19).\\n- The digits had no logical connection to the items' actual value.\",\"prompt\":\"The student wants to highlight the central finding of the experiment. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"Dan Ariely and his colleagues conducted a behavioral economics experiment in which participants bid on items including wine, chocolates, and electronics.\",\"B\":\"Although the last two digits of a Social Security number bear no logical relation to an item's value, participants with higher digit pairs bid up to 346% more than those with lower pairs.\",\"C\":\"In one behavioral economics study, participants were asked to write down the last two digits of their Social Security number before bidding on a series of consumer items.\",\"D\":\"Anchoring is a cognitive bias in which an arbitrary number influences subsequent judgments, as Dan Ariely and colleagues demonstrated in their experiment.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"}]" }, "status": "ok" }, { "attempts": 1, "index": 7, "output": { "costUsd": 0.08304, "model": "opus", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 5 SAT Reading & Writing questions.\nType: vocabulary_in_context | Test: 99 | Unit: 2 | Start Q#: 1\n\nMANDATORY: Your JSON array MUST contain exactly 5 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nFollow the instructions for your type:\n\n[vocabulary_in_context]\n- 2-3 sentence passage with ______ blank from specific academic domains (molecular biology, Renaissance art, behavioral economics, astrophysics, political philosophy)\n- 4 college-level word choices from the same semantic field\n- Correct word is most PRECISE for context, not most common synonym\n- ID format: sat{testNumber}-rw-m{mod}-q{N}\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"vocabulary_in_context\",\"passage\":\"...with ______...\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[text_comprehension]\n- 3-8 sentence passage from published-quality source (journal, novel, essay) with specific names/dates/findings\n- Mix subtypes: main idea (2-3), author's purpose (1-2), inference (2-3), tone (0-1), structure (1-2)\n- Topics: natural science, social science, literature, history, arts/humanities (use each)\n- Distractors: surface-level misreading, wrong scope, opposite interpretation\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"text_comprehension\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[data_analysis]\n- Passage describing realistic data/study results requiring INTERPRETATION, not lookup\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Data patterns: diminishing returns, optimal values, conditional relationships\n- Distractors: wrong column, wrong trend direction, correlation vs causation confusion\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"data_analysis\",\"passage\":\"...\",\"visual\":{...},\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[claim_support]\n- Mix: \"Which finding supports/undermines the claim?\" AND \"Which quotation illustrates the claim?\"\n- Claims must be specific and testable. Distractors are topically related but don't address the specific mechanism.\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"claim_support\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[standard_english_conventions]\n- Passage with [underlined: text here] portion. Student picks best replacement.\n- Cover: subject-verb agreement, pronouns, verb tense, parallel structure, modifiers, commas, semicolons, apostrophes, sentence boundaries\n- \"NO CHANGE\" (choice A) correct ~25% of the time, NOT always\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"standard_english_conventions\",\"passage\":\"...with [underlined: ...]...\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[transitions]\n- Passage with ______ blank connecting two ideas. All 4 choices grammatically correct; distinction is logical.\n- Cover: contrast, addition, cause/effect, example, sequence, concession\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"transitions\",\"passage\":\"...______ ...\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[rhetorical_synthesis]\n- Notes format: \"While researching [topic], a student has taken these notes:\\n- Note 1\\n- Note 2\\n- Note 3\"\n- Specific rhetorical goal stated. All 4 choices are grammatical sentences using the notes; only 1 achieves the goal.\n- Goals: emphasize difference, highlight finding, introduce topic, summarize conclusion, present cause-effect\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"rhetorical_synthesis\",\"passage\":\"...\",\"prompt\":\"The student wants to [goal]. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\nOutput ONLY a valid JSON array of EXACTLY 5 objects. No markdown. Count them before outputting.", "text": "[{\"id\":\"sat99-rw-m2-q1\",\"number\":1,\"type\":\"vocabulary_in_context\",\"passage\":\"In studies of chromatin remodeling, researchers have observed that certain ATP-dependent complexes do not merely displace nucleosomes but ______ them along the DNA strand, repositioning histone octamers with remarkable spatial precision.\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"translocate\",\"B\":\"transport\",\"C\":\"transfer\",\"D\":\"transmit\"},\"answer_format\":\"mcq\",\"correct_answer\":\"A\"},{\"id\":\"sat99-rw-m2-q2\",\"number\":2,\"type\":\"vocabulary_in_context\",\"passage\":\"Art historian Patricia Rubin has argued that Vasari's biographies of Renaissance painters were never intended as neutral records; rather, they ______ a Florentine-centered narrative in which artists outside the Medici orbit were systematically diminished. This editorial shaping, she contends, distorted the historiography of Italian art for centuries.\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"disclosed\",\"B\":\"advanced\",\"C\":\"tolerated\",\"D\":\"inherited\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m2-q3\",\"number\":3,\"type\":\"vocabulary_in_context\",\"passage\":\"In behavioral economics, Richard Thaler's concept of mental accounting suggests that consumers do not treat money as fully ______. A tax refund, a birthday gift, and a paycheck of equal value will be spent quite differently, even though rational choice theory predicts identical treatment.\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"divisible\",\"B\":\"transparent\",\"C\":\"fungible\",\"D\":\"negotiable\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m2-q4\",\"number\":4,\"type\":\"vocabulary_in_context\",\"passage\":\"Observations from the James Webb Space Telescope have revealed galaxies at redshifts exceeding $z=13$ that appear far more luminous and structurally mature than existing cosmological models ______. Theorists are now reconsidering whether early star formation proceeded at rates previously thought impossible.\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"accommodate\",\"B\":\"recognize\",\"C\":\"endorse\",\"D\":\"withstand\"},\"answer_format\":\"mcq\",\"correct_answer\":\"A\"},{\"id\":\"sat99-rw-m2-q5\",\"number\":5,\"type\":\"vocabulary_in_context\",\"passage\":\"In \\\"On Liberty,\\\" John Stuart Mill warns that a citizenry may grow so accustomed to the tyranny of majority opinion that dissent becomes not merely unwelcome but nearly ______. Social pressure, he argued, can silence heterodox thought more effectively than any formal censor.\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"unthinkable\",\"B\":\"indifferent\",\"C\":\"redundant\",\"D\":\"ambivalent\"},\"answer_format\":\"mcq\",\"correct_answer\":\"A\"}]" }, "status": "ok" }, { "attempts": 1, "index": 8, "output": { "costUsd": 0.272115, "model": "opus", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 8 SAT Reading & Writing questions.\nType: text_comprehension | Test: 99 | Unit: 2 | Start Q#: 6\n\nMANDATORY: Your JSON array MUST contain exactly 8 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nFollow the instructions for your type:\n\n[vocabulary_in_context]\n- 2-3 sentence passage with ______ blank from specific academic domains (molecular biology, Renaissance art, behavioral economics, astrophysics, political philosophy)\n- 4 college-level word choices from the same semantic field\n- Correct word is most PRECISE for context, not most common synonym\n- ID format: sat{testNumber}-rw-m{mod}-q{N}\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"vocabulary_in_context\",\"passage\":\"...with ______...\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[text_comprehension]\n- 3-8 sentence passage from published-quality source (journal, novel, essay) with specific names/dates/findings\n- Mix subtypes: main idea (2-3), author's purpose (1-2), inference (2-3), tone (0-1), structure (1-2)\n- Topics: natural science, social science, literature, history, arts/humanities (use each)\n- Distractors: surface-level misreading, wrong scope, opposite interpretation\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"text_comprehension\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[data_analysis]\n- Passage describing realistic data/study results requiring INTERPRETATION, not lookup\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Data patterns: diminishing returns, optimal values, conditional relationships\n- Distractors: wrong column, wrong trend direction, correlation vs causation confusion\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"data_analysis\",\"passage\":\"...\",\"visual\":{...},\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[claim_support]\n- Mix: \"Which finding supports/undermines the claim?\" AND \"Which quotation illustrates the claim?\"\n- Claims must be specific and testable. Distractors are topically related but don't address the specific mechanism.\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"claim_support\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[standard_english_conventions]\n- Passage with [underlined: text here] portion. Student picks best replacement.\n- Cover: subject-verb agreement, pronouns, verb tense, parallel structure, modifiers, commas, semicolons, apostrophes, sentence boundaries\n- \"NO CHANGE\" (choice A) correct ~25% of the time, NOT always\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"standard_english_conventions\",\"passage\":\"...with [underlined: ...]...\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[transitions]\n- Passage with ______ blank connecting two ideas. All 4 choices grammatically correct; distinction is logical.\n- Cover: contrast, addition, cause/effect, example, sequence, concession\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"transitions\",\"passage\":\"...______ ...\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[rhetorical_synthesis]\n- Notes format: \"While researching [topic], a student has taken these notes:\\n- Note 1\\n- Note 2\\n- Note 3\"\n- Specific rhetorical goal stated. All 4 choices are grammatical sentences using the notes; only 1 achieves the goal.\n- Goals: emphasize difference, highlight finding, introduce topic, summarize conclusion, present cause-effect\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"rhetorical_synthesis\",\"passage\":\"...\",\"prompt\":\"The student wants to [goal]. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\nOutput ONLY a valid JSON array of EXACTLY 8 objects. No markdown. Count them before outputting.", "text": "[{\"id\":\"sat99-rw-m2-q6\",\"number\":6,\"type\":\"text_comprehension\",\"passage\":\"In her 2019 study of medieval manuscript illumination, art historian Elina Gertsman challenges the long-standing assumption that blank spaces in devotional texts were merely unfinished work awaiting decoration. Drawing on a corpus of over two hundred Books of Hours, Gertsman argues that voids—empty medallions, unpainted margins, and unfilled initials—were often deliberate theological statements. The absence of image, she contends, invited contemplative engagement with the unrepresentable nature of the divine, functioning as a visual corollary to apophatic theology.\",\"prompt\":\"Which choice best describes the main idea of the text?\",\"choices\":{\"A\":\"Gertsman's analysis of Books of Hours reveals that medieval illuminators frequently left manuscripts unfinished due to economic pressures from wealthy patrons.\",\"B\":\"Gertsman proposes that empty spaces in medieval devotional manuscripts should be understood as intentional theological expressions rather than incomplete craftsmanship.\",\"C\":\"Gertsman demonstrates that apophatic theology originated in the artistic practices of medieval manuscript illuminators rather than in formal philosophical treatises.\",\"D\":\"Gertsman's corpus of two hundred Books of Hours proves that medieval viewers preferred unadorned manuscripts to heavily illuminated ones.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m2-q7\",\"number\":7,\"type\":\"text_comprehension\",\"passage\":\"The following is adapted from Willa Cather's 1918 novel My Ántonia. The narrator, Jim Burden, recalls his arrival on the Nebraska prairie as a child.\\n\\nI had the feeling that the world was left behind, that we had got over the edge of it, and were outside man's jurisdiction. I had never before looked up at the sky when there was not a familiar mountain ridge against it. But this was the complete dome of heaven, all there was of it. Between that earth and that sky I felt erased, blotted out. I did not say my prayers that night: here, I felt, what would be would be.\",\"prompt\":\"Which choice best describes the function of the underlined sentence \\\"Between that earth and that sky I felt erased, blotted out\\\" in the text as a whole?\",\"choices\":{\"A\":\"It introduces a geographic description that the narrator will later contradict with more detailed observations.\",\"B\":\"It conveys the narrator's fear that the prairie's inhabitants will not welcome a newcomer from the East.\",\"C\":\"It captures a sense of self-dissolution that motivates the narrator's subsequent abandonment of a familiar ritual.\",\"D\":\"It criticizes the narrator's earlier assumption that mountains were necessary for a meaningful landscape.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m2-q8\",\"number\":8,\"type\":\"text_comprehension\",\"passage\":\"Economist Esther Duflo's randomized controlled trials in rural Rajasthan produced a counterintuitive finding: offering mothers a small bag of lentils as an incentive for vaccinating their children raised full-immunization rates from 6 percent to 39 percent, while the program's per-child cost actually fell. The lentils, Duflo observed, did not persuade skeptics of vaccine safety; rather, they solved a subtler problem. Many parents already intended to immunize but perpetually postponed the clinic visit. The modest reward supplied the nudge needed to convert latent intention into action.\",\"prompt\":\"Based on the text, what can reasonably be inferred about the parents in Duflo's study?\",\"choices\":{\"A\":\"Most parents were skeptical of vaccine safety and required material incentives to overcome their doubts.\",\"B\":\"Parents in rural Rajasthan generally placed greater value on food staples than on their children's long-term health.\",\"C\":\"Many parents' failure to vaccinate reflected procrastination rather than an objection to immunization itself.\",\"D\":\"The program's success depended primarily on reducing the travel distance between villages and immunization clinics.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m2-q9\",\"number\":9,\"type\":\"text_comprehension\",\"passage\":\"Paleontologist Jingmai O'Connor has argued that the traditional dichotomy between \\\"avian\\\" and \\\"non-avian\\\" dinosaurs obscures more than it clarifies. Her analyses of Cretaceous specimens from the Jehol Biota suggest that features once considered diagnostic of birds—feathers, wishbones, even rudimentary flight—appeared sporadically across multiple theropod lineages. To speak of \\\"the origin of birds,\\\" O'Connor writes, is to impose a retrospective clarity on what was, in the Mesozoic, a thicket of parallel experiments.\",\"prompt\":\"Which choice best describes the author's purpose in the text?\",\"choices\":{\"A\":\"To argue that O'Connor's research has definitively resolved longstanding debates about which theropod lineage gave rise to modern birds.\",\"B\":\"To explain how O'Connor's work complicates a familiar evolutionary narrative by emphasizing the distributed nature of bird-like traits.\",\"C\":\"To criticize paleontologists who continue to use the terms \\\"avian\\\" and \\\"non-avian\\\" despite evidence from the Jehol Biota.\",\"D\":\"To describe a methodological controversy about how fossil specimens from the Cretaceous should be classified and catalogued.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m2-q10\",\"number\":10,\"type\":\"text_comprehension\",\"passage\":\"In a 2021 review of urban heat-island mitigation, civil engineer Hashem Akbari surveyed three decades of rooftop albedo studies. He found that while \\\"cool roofs\\\" painted with highly reflective coatings reliably reduced indoor cooling loads by 10–20 percent in low-rise buildings, their neighborhood-scale effects were far less consistent. In some Mediterranean cities, reflected solar radiation from dense clusters of cool roofs modestly raised street-level temperatures for pedestrians, particularly in narrow canyons where the reflected light was repeatedly absorbed by adjacent walls.\",\"prompt\":\"Which choice best describes the structure of the text?\",\"choices\":{\"A\":\"It presents a widely accepted engineering solution and then describes the mechanism by which it was originally developed.\",\"B\":\"It contrasts two competing theories about urban heat mitigation and argues in favor of the more recent one.\",\"C\":\"It introduces a well-established benefit of a technology and then qualifies that benefit by noting a context-dependent drawback.\",\"D\":\"It summarizes a historical debate and concludes that the debate has been resolved by Akbari's 2021 review.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m2-q11\",\"number\":11,\"type\":\"text_comprehension\",\"passage\":\"The following is adapted from an 1891 essay by Vernon Lee on the aesthetics of music.\\n\\nI do not deny that a sonata may be dissected, its modulations charted, its thematic transformations labeled with scholarly precision. Such labor has its uses. But let us not pretend that the sum of these annotations is the sonata. The listener who emerges from the concert hall transfigured has not been moved by a catalogue of devices; she has been moved by something the catalogue cannot capture, and which the catalogue was never designed to record.\",\"prompt\":\"Which choice best describes the tone of the passage?\",\"choices\":{\"A\":\"Dismissive—Lee rejects musical analysis as a futile endeavor that produces no genuine understanding.\",\"B\":\"Measured—Lee concedes the value of technical analysis while insisting on its insufficiency for explaining musical experience.\",\"C\":\"Nostalgic—Lee laments the decline of an earlier era in which listeners were more emotionally responsive to music.\",\"D\":\"Defensive—Lee justifies her own lack of formal musical training by attacking those who possess it.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m2-q12\",\"number\":12,\"type\":\"text_comprehension\",\"passage\":\"Historians of the early Qing dynasty have long debated the significance of the 1645 queue edict, which required Han Chinese men to adopt the Manchu hairstyle on pain of death. Recent work by Joanna Waley-Cohen reframes the edict as less an assertion of ethnic supremacy than a calculated test of political loyalty. Because the hairstyle was visible at a glance, it allowed Qing officials to identify dissent without the laborious inquisitions that had undermined earlier dynasties. The edict's cruelty, Waley-Cohen suggests, should not obscure its administrative elegance.\",\"prompt\":\"Based on the text, which of the following best describes Waley-Cohen's interpretation of the queue edict?\",\"choices\":{\"A\":\"It was primarily an expression of Manchu cultural chauvinism against Han Chinese traditions.\",\"B\":\"It was a comparatively humane policy when measured against the inquisitions of earlier dynasties.\",\"C\":\"It functioned as an efficient surveillance mechanism that reduced the need for more invasive forms of investigation.\",\"D\":\"It was an impractical policy whose administrative costs outweighed its benefits to the Qing state.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m2-q13\",\"number\":13,\"type\":\"text_comprehension\",\"passage\":\"When the astronomer Vera Rubin began measuring the rotation curves of spiral galaxies in the 1970s, she expected the outer stars to orbit more slowly than those near the galactic center, just as distant planets orbit the sun more slowly than nearby ones. Instead, her spectra showed that stars at the edges of galaxies like Andromeda were moving just as fast as those much closer in. The implication, Rubin recognized, was either that the law of gravity behaved differently on galactic scales or that each galaxy was embedded in a vast halo of unseen mass.\",\"prompt\":\"Which choice best states the main idea of the text?\",\"choices\":{\"A\":\"Rubin's spectra of Andromeda disproved the Newtonian laws of gravity and established a new framework for galactic dynamics.\",\"B\":\"Rubin's observations of galactic rotation contradicted her expectations and pointed toward two possible explanations with profound implications.\",\"C\":\"Rubin's work demonstrated that spiral galaxies rotate differently from solar systems because their stars are more massive than planets.\",\"D\":\"Rubin initially misinterpreted her spectra but later corrected her findings after comparing them with planetary orbital data.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"}]" }, "status": "ok" }, { "attempts": 1, "index": 9, "output": { "costUsd": 0.17019, "model": "opus", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 4 SAT Reading & Writing questions.\nType: data_analysis | Test: 99 | Unit: 2 | Start Q#: 14\n\nMANDATORY: Your JSON array MUST contain exactly 4 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nFollow the instructions for your type:\n\n[vocabulary_in_context]\n- 2-3 sentence passage with ______ blank from specific academic domains (molecular biology, Renaissance art, behavioral economics, astrophysics, political philosophy)\n- 4 college-level word choices from the same semantic field\n- Correct word is most PRECISE for context, not most common synonym\n- ID format: sat{testNumber}-rw-m{mod}-q{N}\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"vocabulary_in_context\",\"passage\":\"...with ______...\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[text_comprehension]\n- 3-8 sentence passage from published-quality source (journal, novel, essay) with specific names/dates/findings\n- Mix subtypes: main idea (2-3), author's purpose (1-2), inference (2-3), tone (0-1), structure (1-2)\n- Topics: natural science, social science, literature, history, arts/humanities (use each)\n- Distractors: surface-level misreading, wrong scope, opposite interpretation\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"text_comprehension\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[data_analysis]\n- Passage describing realistic data/study results requiring INTERPRETATION, not lookup\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Data patterns: diminishing returns, optimal values, conditional relationships\n- Distractors: wrong column, wrong trend direction, correlation vs causation confusion\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"data_analysis\",\"passage\":\"...\",\"visual\":{...},\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[claim_support]\n- Mix: \"Which finding supports/undermines the claim?\" AND \"Which quotation illustrates the claim?\"\n- Claims must be specific and testable. Distractors are topically related but don't address the specific mechanism.\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"claim_support\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[standard_english_conventions]\n- Passage with [underlined: text here] portion. Student picks best replacement.\n- Cover: subject-verb agreement, pronouns, verb tense, parallel structure, modifiers, commas, semicolons, apostrophes, sentence boundaries\n- \"NO CHANGE\" (choice A) correct ~25% of the time, NOT always\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"standard_english_conventions\",\"passage\":\"...with [underlined: ...]...\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[transitions]\n- Passage with ______ blank connecting two ideas. All 4 choices grammatically correct; distinction is logical.\n- Cover: contrast, addition, cause/effect, example, sequence, concession\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"transitions\",\"passage\":\"...______ ...\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[rhetorical_synthesis]\n- Notes format: \"While researching [topic], a student has taken these notes:\\n- Note 1\\n- Note 2\\n- Note 3\"\n- Specific rhetorical goal stated. All 4 choices are grammatical sentences using the notes; only 1 achieves the goal.\n- Goals: emphasize difference, highlight finding, introduce topic, summarize conclusion, present cause-effect\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"rhetorical_synthesis\",\"passage\":\"...\",\"prompt\":\"The student wants to [goal]. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\nOutput ONLY a valid JSON array of EXACTLY 4 objects. No markdown. Count them before outputting.", "text": "[{\"id\":\"sat99-rw-m2-q14\",\"number\":14,\"type\":\"data_analysis\",\"passage\":\"Soil ecologist Marisol Vega investigated whether mycorrhizal fungi inoculation affects the drought tolerance of seedlings in degraded pastureland. She grew seedlings of four native tree species under identical water-restricted conditions, half with inoculation and half without, and measured survival rates after 90 days. Vega hypothesized that the benefit of inoculation would be greatest for species with naturally low drought tolerance, as measured by baseline (non-inoculated) survival.\",\"visual\":{\"type\":\"table\",\"title\":\"90-Day Seedling Survival Rates Under Drought Stress (%)\",\"headers\":[\"Species\",\"Non-inoculated\",\"Inoculated\",\"Change (pp)\"],\"rows\":[[\"Quercus rugosa\",\"72\",\"81\",\"+9\"],[\"Pinus leiophylla\",\"58\",\"74\",\"+16\"],[\"Alnus acuminata\",\"34\",\"62\",\"+28\"],[\"Prunus serotina\",\"49\",\"65\",\"+16\"]]},\"prompt\":\"Which choice best describes data from the table that support Vega's hypothesis?\",\"choices\":{\"A\":\"Alnus acuminata had the lowest non-inoculated survival rate and showed the largest increase in survival following inoculation.\",\"B\":\"Quercus rugosa had the highest non-inoculated survival rate and showed the largest increase in survival following inoculation.\",\"C\":\"Pinus leiophylla and Prunus serotina had identical increases in survival despite different baseline survival rates.\",\"D\":\"Every species showed an increase in survival following inoculation, with the average increase exceeding 15 percentage points.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"A\"},{\"id\":\"sat99-rw-m2-q15\",\"number\":15,\"type\":\"data_analysis\",\"passage\":\"In a study of consumer decision-making, economist Tarun Shah presented participants with a choice between two retirement savings plans that differed only in how contribution rates were framed: Plan X described contributions as a percentage of gross salary, while Plan Y described the same contributions as a fixed dollar amount per paycheck. Shah recorded the proportion of participants in each age group who selected Plan Y.\",\"visual\":{\"type\":\"table\",\"title\":\"Proportion Selecting Plan Y (Fixed-Dollar Framing) by Age Group\",\"headers\":[\"Age Group\",\"Plan Y Selection Rate\",\"Reported Confidence in Choice\"],\"rows\":[[\"18–29\",\"0.71\",\"Low\"],[\"30–44\",\"0.58\",\"Moderate\"],[\"45–59\",\"0.44\",\"Moderate\"],[\"60+\",\"0.29\",\"High\"]]},\"prompt\":\"Based on the table, which statement most accurately characterizes the relationship between age and framing preference?\",\"choices\":{\"A\":\"Older participants were more likely to select the fixed-dollar plan, suggesting increased susceptibility to concrete framing with age.\",\"B\":\"As age increased, preference for the fixed-dollar framing decreased, and reported confidence in the choice increased.\",\"C\":\"Confidence in the choice was lowest among participants who most frequently selected the percentage-based framing.\",\"D\":\"Age had no systematic effect on framing preference, since selection rates varied unpredictably across groups.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m2-q16\",\"number\":16,\"type\":\"data_analysis\",\"passage\":\"To assess whether urban noise affects the vocal behavior of songbirds, ornithologist Kenji Arai recorded dawn choruses of white-crowned sparrows at four sites in and around San Francisco. At each site, he measured the ambient noise level (in decibels) at 5:00 a.m. and the minimum frequency (in hertz) of the sparrows' songs, reasoning that birds in noisier environments might raise their minimum song frequency to avoid masking by low-frequency traffic noise.\",\"visual\":{\"type\":\"table\",\"title\":\"Ambient Noise and Minimum Song Frequency at Four Sites\",\"headers\":[\"Site\",\"Ambient Noise (dB)\",\"Minimum Song Frequency (Hz)\"],\"rows\":[[\"Presidio (suburban edge)\",\"42\",\"2,050\"],[\"Golden Gate Park\",\"51\",\"2,310\"],[\"Mission District\",\"63\",\"2,740\"],[\"SOMA (downtown)\",\"71\",\"3,020\"]]},\"prompt\":\"Which choice best describes data from the table that support Arai's reasoning?\",\"choices\":{\"A\":\"The Presidio site had the lowest ambient noise and the lowest minimum song frequency, while SOMA had the highest values for both.\",\"B\":\"Minimum song frequency at each site was more than 40 times the ambient noise level measured at that site.\",\"C\":\"The difference in minimum song frequency between Golden Gate Park and the Mission District was greater than the difference between the Presidio and Golden Gate Park.\",\"D\":\"Sites with higher ambient noise showed lower minimum song frequencies, contrary to what would be expected if traffic noise masked low-frequency song.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"A\"},{\"id\":\"sat99-rw-m2-q17\",\"number\":17,\"type\":\"data_analysis\",\"passage\":\"Historian Elena Costa examined the circulation numbers of four political pamphlets published in Venice between 1575 and 1590 to investigate whether length influenced popular reach. She hypothesized that shorter pamphlets would achieve wider circulation because they were cheaper to print and easier for semi-literate readers to consume in a single sitting.\",\"visual\":{\"type\":\"table\",\"title\":\"Venetian Political Pamphlets, 1575–1590\",\"headers\":[\"Pamphlet\",\"Length (pages)\",\"Estimated Copies Circulated\",\"Price (soldi)\"],\"rows\":[[\"Discorso sulla Repubblica\",\"48\",\"1,200\",\"12\"],[\"Lettera ai Cittadini\",\"8\",\"9,400\",\"2\"],[\"Trattato della Giustizia\",\"64\",\"850\",\"16\"],[\"Avviso al Popolo\",\"12\",\"7,100\",\"3\"]]},\"prompt\":\"Which choice best describes data from the table that support Costa's hypothesis?\",\"choices\":{\"A\":\"The two longest pamphlets, each exceeding 40 pages, circulated in fewer than 1,500 copies, while the two shortest, under 15 pages, each circulated in more than 7,000 copies.\",\"B\":\"Lettera ai Cittadini circulated in more copies than any other pamphlet, and its price was the lowest of the four.\",\"C\":\"The pamphlet with the highest price, Trattato della Giustizia, circulated in the fewest copies, suggesting that cost rather than length drove readership.\",\"D\":\"Avviso al Popolo was longer than Lettera ai Cittadini but circulated in fewer copies, showing that length had only a modest effect on reach.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"A\"}]" }, "status": "ok" }, { "attempts": 1, "index": 10, "output": { "costUsd": 0.145665, "model": "opus", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 4 SAT Reading & Writing questions.\nType: claim_support | Test: 99 | Unit: 2 | Start Q#: 18\n\nMANDATORY: Your JSON array MUST contain exactly 4 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nFollow the instructions for your type:\n\n[vocabulary_in_context]\n- 2-3 sentence passage with ______ blank from specific academic domains (molecular biology, Renaissance art, behavioral economics, astrophysics, political philosophy)\n- 4 college-level word choices from the same semantic field\n- Correct word is most PRECISE for context, not most common synonym\n- ID format: sat{testNumber}-rw-m{mod}-q{N}\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"vocabulary_in_context\",\"passage\":\"...with ______...\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[text_comprehension]\n- 3-8 sentence passage from published-quality source (journal, novel, essay) with specific names/dates/findings\n- Mix subtypes: main idea (2-3), author's purpose (1-2), inference (2-3), tone (0-1), structure (1-2)\n- Topics: natural science, social science, literature, history, arts/humanities (use each)\n- Distractors: surface-level misreading, wrong scope, opposite interpretation\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"text_comprehension\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[data_analysis]\n- Passage describing realistic data/study results requiring INTERPRETATION, not lookup\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Data patterns: diminishing returns, optimal values, conditional relationships\n- Distractors: wrong column, wrong trend direction, correlation vs causation confusion\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"data_analysis\",\"passage\":\"...\",\"visual\":{...},\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[claim_support]\n- Mix: \"Which finding supports/undermines the claim?\" AND \"Which quotation illustrates the claim?\"\n- Claims must be specific and testable. Distractors are topically related but don't address the specific mechanism.\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"claim_support\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[standard_english_conventions]\n- Passage with [underlined: text here] portion. Student picks best replacement.\n- Cover: subject-verb agreement, pronouns, verb tense, parallel structure, modifiers, commas, semicolons, apostrophes, sentence boundaries\n- \"NO CHANGE\" (choice A) correct ~25% of the time, NOT always\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"standard_english_conventions\",\"passage\":\"...with [underlined: ...]...\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[transitions]\n- Passage with ______ blank connecting two ideas. All 4 choices grammatically correct; distinction is logical.\n- Cover: contrast, addition, cause/effect, example, sequence, concession\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"transitions\",\"passage\":\"...______ ...\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[rhetorical_synthesis]\n- Notes format: \"While researching [topic], a student has taken these notes:\\n- Note 1\\n- Note 2\\n- Note 3\"\n- Specific rhetorical goal stated. All 4 choices are grammatical sentences using the notes; only 1 achieves the goal.\n- Goals: emphasize difference, highlight finding, introduce topic, summarize conclusion, present cause-effect\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"rhetorical_synthesis\",\"passage\":\"...\",\"prompt\":\"The student wants to [goal]. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\nOutput ONLY a valid JSON array of EXACTLY 4 objects. No markdown. Count them before outputting.", "text": "[{\"id\":\"sat99-rw-m2-q18\",\"number\":18,\"type\":\"claim_support\",\"passage\":\"In a 2019 study, marine biologist Elena Vasquez examined the foraging behavior of Caribbean reef octopuses (Octopus briareus) and proposed that these cephalopods develop individualized hunting strategies that persist across multiple foraging trips, rather than responding identically to similar environmental stimuli. Vasquez argued that this consistency reflects stable behavioral differences between individuals analogous to personality traits in vertebrates.\",\"prompt\":\"Which finding from Vasquez's study, if true, would most directly support her claim?\",\"choices\":{\"A\":\"Octopuses in the study showed a marked preference for crustacean prey over mollusks, regardless of which prey type was more abundant in their immediate habitat.\",\"B\":\"When presented with identical artificial reef environments, individual octopuses each used distinct hunting techniques—such as ambush, pursuit, or den-baiting—that each animal reliably repeated across twelve separate trials.\",\"C\":\"Octopuses observed in the wild exhibited more varied hunting behaviors than those kept in controlled laboratory conditions, suggesting that environmental complexity shapes foraging choices.\",\"D\":\"Juvenile octopuses in the study displayed a wider range of hunting techniques than mature adults, whose behaviors became increasingly uniform as they aged.\",\"E\":\"\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m2-q19\",\"number\":19,\"type\":\"claim_support\",\"passage\":\"In her 2021 analysis of Dutch Golden Age still life painting, art historian Margriet de Vries contends that Clara Peeters's signature technique of embedding tiny self-portraits within the reflective surfaces of depicted pewter vessels was not merely a display of technical virtuosity but a deliberate assertion of authorial presence in a genre that typically effaced the painter's identity. De Vries argues that these reflections functioned as visual signatures at a time when female artists were frequently denied formal attribution.\",\"prompt\":\"Which finding, if true, would most strongly support de Vries's claim?\",\"choices\":{\"A\":\"X-ray analysis of several Peeters paintings reveals that the reflected self-portraits were painted in a final, separate layer using pigments distinct from those used in the surrounding still life composition.\",\"B\":\"Peeters's contemporaries, including male still life painters such as Osias Beert, occasionally included reflections of windows or studio interiors in their depicted metalware.\",\"C\":\"Archival records indicate that several of Peeters's still lifes were sold to wealthy patrons in Antwerp without any written attribution to the artist, and that she signed far fewer works than her male contemporaries.\",\"D\":\"Technical examination of pewter vessels from the early seventeenth century shows that their polished surfaces would have produced recognizable mirror-like reflections under studio lighting.\",\"E\":\"\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m2-q20\",\"number\":20,\"type\":\"claim_support\",\"passage\":\"In The Theory of Moral Sentiments (1759), Adam Smith argues that human beings possess an innate capacity for sympathy, which he defines not as pity but as the imaginative act of placing oneself in another's situation. Smith further claims that our moral judgments of others arise primarily from this sympathetic projection rather than from abstract reasoning about duty or consequence.\",\"prompt\":\"Which quotation from The Theory of Moral Sentiments most directly illustrates Smith's claim about the origin of moral judgment?\",\"choices\":{\"A\":\"\\\"Mankind, though naturally sympathetic, never conceive, for what has befallen another, that degree of passion which naturally animates the person principally concerned.\\\"\",\"B\":\"\\\"The man of the most perfect virtue... is he who joins, to the most perfect command of his own original and selfish feelings, the most exquisite sensibility both to the original and sympathetic feelings of others.\\\"\",\"C\":\"\\\"As we have no immediate experience of what other men feel, we can form no idea of the manner in which they are affected, but by conceiving what we ourselves should feel in the like situation.\\\"\",\"D\":\"\\\"Every faculty in one man is the measure by which he judges of the like faculty in another. I judge of your sight by my sight, of your ear by my ear, of your reason by my reason.\\\"\",\"E\":\"\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m2-q21\",\"number\":21,\"type\":\"claim_support\",\"passage\":\"Behavioral economist Priya Narayanan has proposed that so-called \\\"default effects\\\"—the tendency of people to stick with pre-selected options—stem not from genuine preference alignment but from the cognitive effort required to evaluate alternatives. Narayanan predicts that when decision-makers are given more time and cognitive resources to deliberate, the influence of defaults on their final choices will diminish substantially.\",\"prompt\":\"Which finding from a controlled experiment would most directly undermine Narayanan's claim?\",\"choices\":{\"A\":\"Participants who were given a forty-five minute deliberation period and reminded to consider all options chose the default option at essentially the same rate as participants who were required to decide within thirty seconds.\",\"B\":\"Participants under time pressure were more likely to report feeling anxious about their decisions than participants given ample time to deliberate, though both groups selected the default at similar rates.\",\"C\":\"Participants who were required to justify their decisions in writing chose the default less often than those who were not asked to provide justifications, regardless of how much time they were given.\",\"D\":\"Participants who were explicitly told which option had been pre-selected as the default were more likely to choose it than participants who were not informed about the default's status.\",\"E\":\"\"},\"answer_format\":\"mcq\",\"correct_answer\":\"A\"}]" }, "status": "ok" }, { "attempts": 1, "index": 11, "output": { "costUsd": 0.07389, "model": "opus", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 5 SAT Reading & Writing questions.\nType: standard_english_conventions | Test: 99 | Unit: 2 | Start Q#: 22\n\nMANDATORY: Your JSON array MUST contain exactly 5 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nFollow the instructions for your type:\n\n[vocabulary_in_context]\n- 2-3 sentence passage with ______ blank from specific academic domains (molecular biology, Renaissance art, behavioral economics, astrophysics, political philosophy)\n- 4 college-level word choices from the same semantic field\n- Correct word is most PRECISE for context, not most common synonym\n- ID format: sat{testNumber}-rw-m{mod}-q{N}\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"vocabulary_in_context\",\"passage\":\"...with ______...\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[text_comprehension]\n- 3-8 sentence passage from published-quality source (journal, novel, essay) with specific names/dates/findings\n- Mix subtypes: main idea (2-3), author's purpose (1-2), inference (2-3), tone (0-1), structure (1-2)\n- Topics: natural science, social science, literature, history, arts/humanities (use each)\n- Distractors: surface-level misreading, wrong scope, opposite interpretation\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"text_comprehension\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[data_analysis]\n- Passage describing realistic data/study results requiring INTERPRETATION, not lookup\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Data patterns: diminishing returns, optimal values, conditional relationships\n- Distractors: wrong column, wrong trend direction, correlation vs causation confusion\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"data_analysis\",\"passage\":\"...\",\"visual\":{...},\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[claim_support]\n- Mix: \"Which finding supports/undermines the claim?\" AND \"Which quotation illustrates the claim?\"\n- Claims must be specific and testable. Distractors are topically related but don't address the specific mechanism.\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"claim_support\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[standard_english_conventions]\n- Passage with [underlined: text here] portion. Student picks best replacement.\n- Cover: subject-verb agreement, pronouns, verb tense, parallel structure, modifiers, commas, semicolons, apostrophes, sentence boundaries\n- \"NO CHANGE\" (choice A) correct ~25% of the time, NOT always\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"standard_english_conventions\",\"passage\":\"...with [underlined: ...]...\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[transitions]\n- Passage with ______ blank connecting two ideas. All 4 choices grammatically correct; distinction is logical.\n- Cover: contrast, addition, cause/effect, example, sequence, concession\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"transitions\",\"passage\":\"...______ ...\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[rhetorical_synthesis]\n- Notes format: \"While researching [topic], a student has taken these notes:\\n- Note 1\\n- Note 2\\n- Note 3\"\n- Specific rhetorical goal stated. All 4 choices are grammatical sentences using the notes; only 1 achieves the goal.\n- Goals: emphasize difference, highlight finding, introduce topic, summarize conclusion, present cause-effect\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"rhetorical_synthesis\",\"passage\":\"...\",\"prompt\":\"The student wants to [goal]. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\nOutput ONLY a valid JSON array of EXACTLY 5 objects. No markdown. Count them before outputting.", "text": "[{\"id\":\"sat99-rw-m2-q22\",\"number\":22,\"type\":\"standard_english_conventions\",\"passage\":\"The archaeologist Kathleen Kenyon, whose excavations at Jericho in the 1950s revolutionized the field's understanding of Neolithic settlement, [underlined: are] now regarded as a pioneer of stratigraphic methodology.\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"is\",\"C\":\"were\",\"D\":\"being\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m2-q23\",\"number\":23,\"type\":\"standard_english_conventions\",\"passage\":\"Marine biologists studying the Great Barrier Reef have documented a surprising resilience in certain coral species; [underlined: however these] organisms continue to reproduce even after repeated bleaching events.\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"however, these\",\"C\":\"however: these\",\"D\":\"however these,\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m2-q24\",\"number\":24,\"type\":\"standard_english_conventions\",\"passage\":\"When composing her final symphony, Florence Price drew on spirituals, European classical forms, and [underlined: she incorporated syncopated rhythms] characteristic of African American folk traditions.\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"incorporating syncopated rhythms\",\"C\":\"syncopated rhythms\",\"D\":\"syncopated rhythms were incorporated\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m2-q25\",\"number\":25,\"type\":\"standard_english_conventions\",\"passage\":\"The committee's decision to fund the observatory surprised many researchers [underlined: who's] grant proposals had been rejected in earlier rounds of review.\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"whose\",\"C\":\"whos'\",\"D\":\"whom's\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m2-q26\",\"number\":26,\"type\":\"standard_english_conventions\",\"passage\":\"By the time the expedition reached the base camp on the southern glacier, the team [underlined: has endured] three weeks of subzero temperatures and persistent whiteout conditions.\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"endures\",\"C\":\"is enduring\",\"D\":\"had endured\"},\"answer_format\":\"mcq\",\"correct_answer\":\"D\"}]" }, "status": "ok" }, { "attempts": 1, "index": 12, "output": { "costUsd": 0.082365, "model": "opus", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 4 SAT Reading & Writing questions.\nType: transitions | Test: 99 | Unit: 2 | Start Q#: 27\n\nMANDATORY: Your JSON array MUST contain exactly 4 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nFollow the instructions for your type:\n\n[vocabulary_in_context]\n- 2-3 sentence passage with ______ blank from specific academic domains (molecular biology, Renaissance art, behavioral economics, astrophysics, political philosophy)\n- 4 college-level word choices from the same semantic field\n- Correct word is most PRECISE for context, not most common synonym\n- ID format: sat{testNumber}-rw-m{mod}-q{N}\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"vocabulary_in_context\",\"passage\":\"...with ______...\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[text_comprehension]\n- 3-8 sentence passage from published-quality source (journal, novel, essay) with specific names/dates/findings\n- Mix subtypes: main idea (2-3), author's purpose (1-2), inference (2-3), tone (0-1), structure (1-2)\n- Topics: natural science, social science, literature, history, arts/humanities (use each)\n- Distractors: surface-level misreading, wrong scope, opposite interpretation\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"text_comprehension\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[data_analysis]\n- Passage describing realistic data/study results requiring INTERPRETATION, not lookup\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Data patterns: diminishing returns, optimal values, conditional relationships\n- Distractors: wrong column, wrong trend direction, correlation vs causation confusion\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"data_analysis\",\"passage\":\"...\",\"visual\":{...},\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[claim_support]\n- Mix: \"Which finding supports/undermines the claim?\" AND \"Which quotation illustrates the claim?\"\n- Claims must be specific and testable. Distractors are topically related but don't address the specific mechanism.\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"claim_support\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[standard_english_conventions]\n- Passage with [underlined: text here] portion. Student picks best replacement.\n- Cover: subject-verb agreement, pronouns, verb tense, parallel structure, modifiers, commas, semicolons, apostrophes, sentence boundaries\n- \"NO CHANGE\" (choice A) correct ~25% of the time, NOT always\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"standard_english_conventions\",\"passage\":\"...with [underlined: ...]...\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[transitions]\n- Passage with ______ blank connecting two ideas. All 4 choices grammatically correct; distinction is logical.\n- Cover: contrast, addition, cause/effect, example, sequence, concession\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"transitions\",\"passage\":\"...______ ...\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[rhetorical_synthesis]\n- Notes format: \"While researching [topic], a student has taken these notes:\\n- Note 1\\n- Note 2\\n- Note 3\"\n- Specific rhetorical goal stated. All 4 choices are grammatical sentences using the notes; only 1 achieves the goal.\n- Goals: emphasize difference, highlight finding, introduce topic, summarize conclusion, present cause-effect\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"rhetorical_synthesis\",\"passage\":\"...\",\"prompt\":\"The student wants to [goal]. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\nOutput ONLY a valid JSON array of EXACTLY 4 objects. No markdown. Count them before outputting.", "text": "[{\"id\":\"sat99-rw-m2-q27\",\"number\":27,\"type\":\"transitions\",\"passage\":\"The axolotl, a salamander native to lakes near Mexico City, retains its larval features—including external gills and a finned tail—throughout adulthood, a phenomenon known as neoteny. Most amphibians undergo metamorphosis to develop lungs and adapt to terrestrial life. ______ the axolotl remains fully aquatic, breeding and feeding in the same form it had as a juvenile.\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"By contrast,\",\"B\":\"Likewise,\",\"C\":\"As a result,\",\"D\":\"For instance,\"},\"answer_format\":\"mcq\",\"correct_answer\":\"A\"},{\"id\":\"sat99-rw-m2-q28\",\"number\":28,\"type\":\"transitions\",\"passage\":\"In her 2019 study of urban tree canopies, ecologist Mariana Velez found that neighborhoods with greater tree cover recorded summer temperatures up to 4°C lower than adjacent treeless blocks. The cooling effect persisted even during prolonged heat waves, when air-conditioning demand spiked elsewhere. ______ Velez concluded that municipal tree-planting initiatives could meaningfully reduce both energy consumption and heat-related illness in dense city centers.\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"Nevertheless,\",\"B\":\"On these grounds,\",\"C\":\"In addition,\",\"D\":\"Granted,\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m2-q29\",\"number\":29,\"type\":\"transitions\",\"passage\":\"Composer Florence Price submitted her Symphony in E minor to the 1932 Wanamaker Foundation competition and won first prize, prompting the Chicago Symphony Orchestra to perform the work the following year. For decades afterward, however, her scores were largely forgotten and her name absent from concert programs. ______ in 2009, a renovation crew discovered a trove of Price's manuscripts in her abandoned summer home, sparking renewed scholarly and performance interest.\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"Consequently,\",\"B\":\"Similarly,\",\"C\":\"Then,\",\"D\":\"Meanwhile,\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m2-q30\",\"number\":30,\"type\":\"transitions\",\"passage\":\"Behavioral economist Sendhil Mullainathan has argued that scarcity—whether of time, money, or food—captures the mind, narrowing attention to the immediate shortage and degrading performance on unrelated tasks. Critics initially dismissed the claim as overstated, suggesting that observed deficits stemmed from chronic stress rather than scarcity itself. ______ subsequent experiments in which participants were temporarily induced to feel scarce, and then released from that state, replicated the cognitive effects within a single laboratory session.\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"Moreover,\",\"B\":\"For example,\",\"C\":\"In short,\",\"D\":\"However,\"},\"answer_format\":\"mcq\",\"correct_answer\":\"D\"}]" }, "status": "ok" }, { "attempts": 1, "index": 13, "output": { "costUsd": 0.12909, "model": "opus", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 3 SAT Reading & Writing questions.\nType: rhetorical_synthesis | Test: 99 | Unit: 2 | Start Q#: 31\n\nMANDATORY: Your JSON array MUST contain exactly 3 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nFollow the instructions for your type:\n\n[vocabulary_in_context]\n- 2-3 sentence passage with ______ blank from specific academic domains (molecular biology, Renaissance art, behavioral economics, astrophysics, political philosophy)\n- 4 college-level word choices from the same semantic field\n- Correct word is most PRECISE for context, not most common synonym\n- ID format: sat{testNumber}-rw-m{mod}-q{N}\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"vocabulary_in_context\",\"passage\":\"...with ______...\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[text_comprehension]\n- 3-8 sentence passage from published-quality source (journal, novel, essay) with specific names/dates/findings\n- Mix subtypes: main idea (2-3), author's purpose (1-2), inference (2-3), tone (0-1), structure (1-2)\n- Topics: natural science, social science, literature, history, arts/humanities (use each)\n- Distractors: surface-level misreading, wrong scope, opposite interpretation\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"text_comprehension\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[data_analysis]\n- Passage describing realistic data/study results requiring INTERPRETATION, not lookup\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Data patterns: diminishing returns, optimal values, conditional relationships\n- Distractors: wrong column, wrong trend direction, correlation vs causation confusion\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"data_analysis\",\"passage\":\"...\",\"visual\":{...},\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[claim_support]\n- Mix: \"Which finding supports/undermines the claim?\" AND \"Which quotation illustrates the claim?\"\n- Claims must be specific and testable. Distractors are topically related but don't address the specific mechanism.\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"claim_support\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[standard_english_conventions]\n- Passage with [underlined: text here] portion. Student picks best replacement.\n- Cover: subject-verb agreement, pronouns, verb tense, parallel structure, modifiers, commas, semicolons, apostrophes, sentence boundaries\n- \"NO CHANGE\" (choice A) correct ~25% of the time, NOT always\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"standard_english_conventions\",\"passage\":\"...with [underlined: ...]...\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[transitions]\n- Passage with ______ blank connecting two ideas. All 4 choices grammatically correct; distinction is logical.\n- Cover: contrast, addition, cause/effect, example, sequence, concession\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"transitions\",\"passage\":\"...______ ...\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[rhetorical_synthesis]\n- Notes format: \"While researching [topic], a student has taken these notes:\\n- Note 1\\n- Note 2\\n- Note 3\"\n- Specific rhetorical goal stated. All 4 choices are grammatical sentences using the notes; only 1 achieves the goal.\n- Goals: emphasize difference, highlight finding, introduce topic, summarize conclusion, present cause-effect\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"rhetorical_synthesis\",\"passage\":\"...\",\"prompt\":\"The student wants to [goal]. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\nOutput ONLY a valid JSON array of EXACTLY 3 objects. No markdown. Count them before outputting.", "text": "[{\"id\":\"sat99-rw-m2-q31\",\"number\":31,\"type\":\"rhetorical_synthesis\",\"passage\":\"While researching the foraging behavior of bumblebees, a student has taken the following notes:\\n- Bumblebees (genus $Bombus$) visit a wide variety of flowering plants.\\n- A 2019 study by Lars Chittka observed bumblebees in an enclosed arena with artificial flowers.\\n- Some flowers offered high-concentration sucrose solutions; others offered low-concentration solutions.\\n- Bumblebees in the study revisited high-concentration flowers significantly more often than low-concentration ones.\\n- Chittka concluded that bumblebees can distinguish between reward magnitudes and adjust foraging accordingly.\",\"prompt\":\"The student wants to present the main finding of Chittka's 2019 study. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"Bumblebees, members of the genus $Bombus$, are known to visit a wide variety of flowering plants in natural environments.\",\"B\":\"In a 2019 study, Lars Chittka observed bumblebees in an enclosed arena containing artificial flowers with varying sucrose concentrations.\",\"C\":\"In Chittka's 2019 study, bumblebees revisited artificial flowers offering high-concentration sucrose solutions significantly more often than those offering low-concentration solutions.\",\"D\":\"Chittka's research suggests that the foraging behavior of bumblebees is shaped by their ability to perceive features of their environment.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m2-q32\",\"number\":32,\"type\":\"rhetorical_synthesis\",\"passage\":\"While researching the Italian Renaissance painter Sofonisba Anguissola, a student has taken the following notes:\\n- Sofonisba Anguissola (c. 1532–1625) was a Northern Italian painter from Cremona.\\n- She trained under Bernardino Campi, an unusual arrangement for women of her era.\\n- Her 1555 painting $The$ $Chess$ $Game$ depicts her three sisters playing chess.\\n- $The$ $Chess$ $Game$ is considered an early example of a genre scene featuring identifiable individuals.\\n- Most genre scenes of the period depicted anonymous, generalized figures rather than specific people.\",\"prompt\":\"The student wants to emphasize a way in which Anguissola's $The$ $Chess$ $Game$ differed from typical genre scenes of its time. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"Sofonisba Anguissola, who lived from approximately 1532 to 1625, was a painter from the Northern Italian city of Cremona.\",\"B\":\"Whereas most genre scenes of the period depicted anonymous, generalized figures, Anguissola's 1555 painting $The$ $Chess$ $Game$ portrayed her three sisters as identifiable individuals.\",\"C\":\"Anguissola trained under the painter Bernardino Campi, an arrangement that was unusual for women during the Italian Renaissance.\",\"D\":\"Anguissola's $The$ $Chess$ $Game$, completed in 1555, depicts her three sisters engaged in a game of chess.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m2-q33\",\"number\":33,\"type\":\"rhetorical_synthesis\",\"passage\":\"While researching the formation of Saturn's rings, a student has taken the following notes:\\n- Saturn's rings are composed primarily of water ice particles ranging from micrometers to meters in size.\\n- One hypothesis proposes the rings formed roughly 4.5 billion years ago, alongside Saturn itself.\\n- A 2019 analysis of Cassini spacecraft data by Iess and colleagues estimated the rings' total mass.\\n- Based on the low mass and minimal contamination by interplanetary dust, Iess and colleagues concluded the rings are likely only 10 to 100 million years old.\\n- This estimate suggests the rings may have formed long after the planet itself, possibly from the breakup of an icy moon.\",\"prompt\":\"The student wants to introduce the topic of the research to an audience unfamiliar with Saturn's rings. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"Saturn's rings, which are composed primarily of water ice particles varying widely in size, have been the subject of recent scientific investigation into their age and origin.\",\"B\":\"Iess and colleagues analyzed Cassini spacecraft data in 2019 to estimate the total mass of Saturn's rings.\",\"C\":\"The low mass and minimal dust contamination of Saturn's rings led Iess and colleagues to conclude that the rings are 10 to 100 million years old.\",\"D\":\"Saturn's rings may have formed from the breakup of an icy moon long after Saturn itself coalesced roughly 4.5 billion years ago.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"A\"}]" }, "status": "ok" } ], "succeeded": 14 }, "gen-math": { "count": 10, "costUsd": 0, "failed": 0, "items": [ { "costUsd": 0, "model": "gemini-2.5-pro", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 7 SAT Math questions (of which 2 are student-produced-response and the rest are MCQ).\nType: algebra | Test: 99 | Unit: 1 | Start Q#: 1\n\nMANDATORY: Your JSON array MUST contain exactly 7 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nMCQ Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"algebra\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\nSPR Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"algebra\",\"prompt\":\"...\",\"answer_format\":\"student_produced\",\"correct_answer\":\"numeric_value\"}\nPlace SPR questions at the END of the array.\n\nFollow instructions for your type:\n\n[algebra] (MCQ + 2 SPR)\n- Topics: linear equations (1 var, 2 vars), systems of 2 linear equations, linear inequalities, absolute value\n- Mix word problems and abstract algebra. Use $...$ notation.\n- MCQ distractors: sign errors, arithmetic mistakes, setup errors, partial answers\n- SPR answers: clean integers or simple fractions\n\n[functions_analysis] (MCQ + 2 SPR)\n- Topics: function evaluation, domain/range, transformations (f(x)+k, f(x-h)), graphical analysis, composition f(g(x)), piecewise\n- At least one conceptual insight question (e.g., (p,q) and (q,p) on same graph)\n- Describe graphs in text. All computations yield clean results.\n\n[geometry_trigonometry] (MCQ + 2 SPR)\n- Topics: circle equations, Pythagorean theorem, special right triangles, coordinate geometry, trig ratios, volume/surface area, angle relationships\n- Optional \"visual\" field: {\"type\":\"coordinate_plane\",\"description\":\"...\"}\n- Radical answers OK (e.g., $5\\sqrt{13}$). Multi-step reasoning required.\n\n[statistics_probability] (MCQ + 2 SPR)\n- Topics: mean/median/mode, standard deviation (conceptual), probability (conditional from tables), data interpretation, scatter plots/line of best fit, sampling/bias\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Distractors: P(A|B) vs P(B|A) confusion, mean vs median, causal from correlational\n\n[advanced_math] (MCQ + 2 SPR)\n- Topics: quadratics (factoring, discriminant), polynomials (zeros, end behavior), exponentials (growth/decay), rational expressions, radical expressions, non-linear systems\n- Hardest questions on the test. Multi-step reasoning required. Elegant algebraic insight preferred.\n- All expressions must be factorable with integer coefficients. SPR answers are clean numbers.\n\nVERIFY EVERY ANSWER: Solve each problem step-by-step before finalizing correct_answer.\nOutput ONLY a valid JSON array of EXACTLY 7 objects. No markdown. Count them before outputting.", "text": "```json\n[\n {\n \"id\": \"sat99-math-m1-q1\",\n \"number\": 1,\n \"type\": \"algebra\",\n \"prompt\": \"A moving company charges a flat fee of $120 for using their truck and an additional $45 per hour for labor. If a customer's final bill was $390, how many hours of labor did the customer pay for?\",\n \"choices\": {\n \"A\": \"2.6\",\n \"B\": \"5\",\n \"C\": \"6\",\n \"D\": \"8.6\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"C\"\n },\n {\n \"id\": \"sat99-math-m1-q2\",\n \"number\": 2,\n \"type\": \"algebra\",\n \"prompt\": \"At a school fundraiser, adults' tickets were sold for $8 and students' tickets were sold for $5. A total of 150 tickets were sold, bringing in a total of $990. How many adult tickets were sold?\",\n \"choices\": {\n \"A\": \"70\",\n \"B\": \"80\",\n \"C\": \"90\",\n \"D\": \"100\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"B\"\n },\n {\n \"id\": \"sat99-math-m1-q3\",\n \"number\": 3,\n \"type\": \"algebra\",\n \"prompt\": \"The atmospheric temperature, $T$, in degrees Celsius, required for a certain scientific experiment must satisfy the inequality $|T - (-20)| \\\\le 1.5$. Which of the following represents all possible values for the temperature $T$?\",\n \"choices\": {\n \"A\": \"T \\\\ge -18.5\",\n \"B\": \"T \\\\le -21.5\",\n \"C\": \"-18.5 \\\\le T \\\\le 18.5\",\n \"D\": \"-21.5 \\\\le T \\\\le -18.5\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"D\"\n },\n {\n \"id\": \"sat99-math-m1-q4\",\n \"number\": 4,\n \"type\": \"algebra\",\n \"prompt\": \"The total cost, $C$, in dollars, for an online subscription is given by the equation $C = 15m + 20$, where $m$ is the number of months. What is the best interpretation of the number 15 in this context?\",\n \"choices\": {\n \"A\": \"The cost per month\",\n \"B\": \"The initial sign-up fee\",\n \"C\": \"The total cost for 20 months\",\n \"D\": \"The number of months included in the subscription\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"A\"\n },\n {\n \"id\": \"sat99-math-m1-q5\",\n \"number\": 5,\n \"type\": \"algebra\",\n \"prompt\": \"If $2x - 3y = 8$ and $x + y = 9$, what is the value of the expression $x - 4y$?\",\n \"choices\": {\n \"A\": \"17\",\n \"B\": \"-1\",\n \"C\": \"1\",\n \"D\": \"7\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"B\"\n },\n {\n \"id\": \"sat99-math-m1-q6\",\n \"number\": 6,\n \"type\": \"algebra\",\n \"prompt\": \"For a certain chemical process, the concentration $C$ of a substance must be maintained such that $|3C - 11| = 4$. What is one possible value for the concentration $C$?\",\n \"answer_format\": \"student_produced\",\n \"correct_answer\": \"5\"\n },\n {\n \"id\": \"sat99-math-m1-q7\",\n \"number\": 7,\n \"type\": \"algebra\",\n \"prompt\": \"If $x > 0$ and $5x - 2y = 15$ and $4x - y = 12$, what is the value of $x$?\",\n \"answer_format\": \"student_produced\",\n \"correct_answer\": \"3\"\n }\n]\n```" }, { "costUsd": 0, "model": "gemini-2.5-pro", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 5 SAT Math questions (of which 1 are student-produced-response and the rest are MCQ).\nType: functions_analysis | Test: 99 | Unit: 1 | Start Q#: 8\n\nMANDATORY: Your JSON array MUST contain exactly 5 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nMCQ Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"functions_analysis\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\nSPR Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"functions_analysis\",\"prompt\":\"...\",\"answer_format\":\"student_produced\",\"correct_answer\":\"numeric_value\"}\nPlace SPR questions at the END of the array.\n\nFollow instructions for your type:\n\n[algebra] (MCQ + 1 SPR)\n- Topics: linear equations (1 var, 2 vars), systems of 2 linear equations, linear inequalities, absolute value\n- Mix word problems and abstract algebra. Use $...$ notation.\n- MCQ distractors: sign errors, arithmetic mistakes, setup errors, partial answers\n- SPR answers: clean integers or simple fractions\n\n[functions_analysis] (MCQ + 1 SPR)\n- Topics: function evaluation, domain/range, transformations (f(x)+k, f(x-h)), graphical analysis, composition f(g(x)), piecewise\n- At least one conceptual insight question (e.g., (p,q) and (q,p) on same graph)\n- Describe graphs in text. All computations yield clean results.\n\n[geometry_trigonometry] (MCQ + 1 SPR)\n- Topics: circle equations, Pythagorean theorem, special right triangles, coordinate geometry, trig ratios, volume/surface area, angle relationships\n- Optional \"visual\" field: {\"type\":\"coordinate_plane\",\"description\":\"...\"}\n- Radical answers OK (e.g., $5\\sqrt{13}$). Multi-step reasoning required.\n\n[statistics_probability] (MCQ + 1 SPR)\n- Topics: mean/median/mode, standard deviation (conceptual), probability (conditional from tables), data interpretation, scatter plots/line of best fit, sampling/bias\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Distractors: P(A|B) vs P(B|A) confusion, mean vs median, causal from correlational\n\n[advanced_math] (MCQ + 1 SPR)\n- Topics: quadratics (factoring, discriminant), polynomials (zeros, end behavior), exponentials (growth/decay), rational expressions, radical expressions, non-linear systems\n- Hardest questions on the test. Multi-step reasoning required. Elegant algebraic insight preferred.\n- All expressions must be factorable with integer coefficients. SPR answers are clean numbers.\n\nVERIFY EVERY ANSWER: Solve each problem step-by-step before finalizing correct_answer.\nOutput ONLY a valid JSON array of EXACTLY 5 objects. No markdown. Count them before outputting.", "text": "```json\n[\n {\n \"id\": \"sat99-math-m1-q8\",\n \"number\": 8,\n \"type\": \"functions_analysis\",\n \"prompt\": \"Let the function $f$ be defined as $f(x) = x^2 + 5$ and the function $g$ be defined as $g(x) = x - 2$. What is the value of $f(g(3))$?\",\n \"choices\": {\n \"A\": \"6\",\n \"B\": \"8\",\n \"C\": \"14\",\n \"D\": \"17\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"A\"\n },\n {\n \"id\": \"sat99-math-m1-q9\",\n \"number\": 9,\n \"type\": \"functions_analysis\",\n \"prompt\": \"The graph of $y = h(x)$ is shown in the $xy$-plane. A new function, $k(x)$, is defined as $k(x) = h(x) - 2$. What is the maximum value of the function $k(x)$?\",\n \"visual\": {\n \"type\": \"coordinate_plane\",\n \"description\": \"A smooth curve representing the function h(x). The curve starts at (-4, -1), goes up to a maximum point at (-2, 5), goes down to a minimum point at (1, 0), and then goes up to the point (3, 3).\"\n },\n \"choices\": {\n \"A\": \"-2\",\n \"B\": \"3\",\n \"C\": \"5\",\n \"D\": \"7\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"B\"\n },\n {\n \"id\": \"sat99-math-m1-q10\",\n \"number\": 10,\n \"type\": \"functions_analysis\",\n \"prompt\": \"A function $f(x)$ has a domain of $x \\ge 2$ and a range of $y \\le 5$. The function $g(x)$ is defined as $g(x) = f(x+3) - 4$. Which of the following represents the domain and range of $g(x)$?\",\n \"choices\": {\n \"A\": \"Domain: $x \\ge 5$, Range: $y \\le 1$\",\n \"B\": \"Domain: $x \\ge -1$, Range: $y \\le 9$\",\n \"C\": \"Domain: $x \\ge 5$, Range: $y \\le 9$\",\n \"D\": \"Domain: $x \\ge -1$, Range: $y \\le 1$\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"D\"\n },\n {\n \"id\": \"sat99-math-m1-q11\",\n \"number\": 11,\n \"type\": \"functions_analysis\",\n \"prompt\": \"For a certain function $f$, it is known that if the point $(a, b)$ is on the graph of $y=f(x)$, then the point $(b, a)$ is also on the graph. If $f(2) = 5$, what is the value of $f(f(2))$?\",\n \"choices\": {\n \"A\": \"0\",\n \"B\": \"2\",\n \"C\": \"5\",\n \"D\": \"The value cannot be determined from the information given.\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"B\"\n },\n {\n \"id\": \"sat99-math-m1-q12\",\n \"number\": 12,\n \"type\": \"functions_analysis\",\n \"prompt\": \"For the piecewise function $g(x)$ defined below,\\n$g(x) = \\\\begin{cases} 4x-3 & \\\\text{if } x < 2 \\\\\\\\ x^2+1 & \\\\text{if } x \\\\ge 2 \\\\end{cases}$\\nWhat is the value of $g(g(1))$?\",\n \"answer_format\": \"student_produced\",\n \"correct_answer\": \"2\"\n }\n]\n```" }, { "costUsd": 0, "model": "gemini-2.5-pro", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 5 SAT Math questions (of which 1 are student-produced-response and the rest are MCQ).\nType: geometry_trigonometry | Test: 99 | Unit: 1 | Start Q#: 13\n\nMANDATORY: Your JSON array MUST contain exactly 5 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nMCQ Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"geometry_trigonometry\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\nSPR Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"geometry_trigonometry\",\"prompt\":\"...\",\"answer_format\":\"student_produced\",\"correct_answer\":\"numeric_value\"}\nPlace SPR questions at the END of the array.\n\nFollow instructions for your type:\n\n[algebra] (MCQ + 1 SPR)\n- Topics: linear equations (1 var, 2 vars), systems of 2 linear equations, linear inequalities, absolute value\n- Mix word problems and abstract algebra. Use $...$ notation.\n- MCQ distractors: sign errors, arithmetic mistakes, setup errors, partial answers\n- SPR answers: clean integers or simple fractions\n\n[functions_analysis] (MCQ + 1 SPR)\n- Topics: function evaluation, domain/range, transformations (f(x)+k, f(x-h)), graphical analysis, composition f(g(x)), piecewise\n- At least one conceptual insight question (e.g., (p,q) and (q,p) on same graph)\n- Describe graphs in text. All computations yield clean results.\n\n[geometry_trigonometry] (MCQ + 1 SPR)\n- Topics: circle equations, Pythagorean theorem, special right triangles, coordinate geometry, trig ratios, volume/surface area, angle relationships\n- Optional \"visual\" field: {\"type\":\"coordinate_plane\",\"description\":\"...\"}\n- Radical answers OK (e.g., $5\\sqrt{13}$). Multi-step reasoning required.\n\n[statistics_probability] (MCQ + 1 SPR)\n- Topics: mean/median/mode, standard deviation (conceptual), probability (conditional from tables), data interpretation, scatter plots/line of best fit, sampling/bias\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Distractors: P(A|B) vs P(B|A) confusion, mean vs median, causal from correlational\n\n[advanced_math] (MCQ + 1 SPR)\n- Topics: quadratics (factoring, discriminant), polynomials (zeros, end behavior), exponentials (growth/decay), rational expressions, radical expressions, non-linear systems\n- Hardest questions on the test. Multi-step reasoning required. Elegant algebraic insight preferred.\n- All expressions must be factorable with integer coefficients. SPR answers are clean numbers.\n\nVERIFY EVERY ANSWER: Solve each problem step-by-step before finalizing correct_answer.\nOutput ONLY a valid JSON array of EXACTLY 5 objects. No markdown. Count them before outputting.", "text": "```json\n[\n {\n \"id\": \"sat99-math-m1-q13\",\n \"number\": 13,\n \"type\": \"geometry_trigonometry\",\n \"prompt\": \"The equation of a circle in the $xy$-plane is $x^2 + y^2 - 6x + 8y + 21 = 0$. What is the radius of the circle?\",\n \"choices\": {\n \"A\": \"5\",\n \"B\": \"$\\\\sqrt{21}$\",\n \"C\": \"2\",\n \"D\": \"4\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"C\"\n },\n {\n \"id\": \"sat99-math-m1-q14\",\n \"number\": 14,\n \"type\": \"geometry_trigonometry\",\n \"prompt\": \"In the $xy$-plane, a line segment is drawn from the origin to a point $P$ in Quadrant I. If the length of the segment is 8 and it makes an angle $\\\\theta$ with the positive $x$-axis such that $\\\\tan\\\\theta = 1$, what is the $y$-coordinate of point $P$?\",\n \"choices\": {\n \"A\": \"4\",\n \"B\": \"$4\\\\sqrt{2}$\",\n \"C\": \"8\",\n \"D\": \"$8\\\\sqrt{2}$\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"B\"\n },\n {\n \"id\": \"sat99-math-m1-q15\",\n \"number\": 15,\n \"type\": \"geometry_trigonometry\",\n \"prompt\": \"In right triangle $XYZ$, the measure of angle $Z$ is $90^{\\\\circ}$. If $\\\\sin(X) = \\\\frac{7}{25}$, what is the value of $\\\\cos(Y)$?\",\n \"choices\": {\n \"A\": \"$\\\\frac{24}{25}$\",\n \"B\": \"$\\\\frac{7}{24}$\",\n \"C\": \"$\\\\frac{25}{7}$\",\n \"D\": \"$\\\\frac{7}{25}$\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"D\"\n },\n {\n \"id\": \"sat99-math-m1-q16\",\n \"number\": 16,\n \"type\": \"geometry_trigonometry\",\n \"prompt\": \"A right circular cylinder is inscribed within a cube. The volume of the cube is 64 cubic centimeters. What is the volume of the cylinder, in cubic centimeters?\",\n \"choices\": {\n \"A\": \"$8\\\\pi$\",\n \"B\": \"$16\\\\pi$\",\n \"C\": \"$32\\\\pi$\",\n \"D\": \"$64\\\\pi$\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"B\"\n },\n {\n \"id\": \"sat99-math-m1-q17\",\n \"number\": 17,\n \"type\": \"geometry_trigonometry\",\n \"prompt\": \"A circle has a radius of 10. The area of a sector of the circle is $10\\\\pi$. What is the measure of the sector's central angle, in degrees?\",\n \"answer_format\": \"student_produced\",\n \"correct_answer\": \"36\"\n }\n]\n```" }, { "costUsd": 0, "model": "gemini-2.5-pro", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 5 SAT Math questions (of which 1 are student-produced-response and the rest are MCQ).\nType: statistics_probability | Test: 99 | Unit: 1 | Start Q#: 18\n\nMANDATORY: Your JSON array MUST contain exactly 5 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nMCQ Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"statistics_probability\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\nSPR Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"statistics_probability\",\"prompt\":\"...\",\"answer_format\":\"student_produced\",\"correct_answer\":\"numeric_value\"}\nPlace SPR questions at the END of the array.\n\nFollow instructions for your type:\n\n[algebra] (MCQ + 1 SPR)\n- Topics: linear equations (1 var, 2 vars), systems of 2 linear equations, linear inequalities, absolute value\n- Mix word problems and abstract algebra. Use $...$ notation.\n- MCQ distractors: sign errors, arithmetic mistakes, setup errors, partial answers\n- SPR answers: clean integers or simple fractions\n\n[functions_analysis] (MCQ + 1 SPR)\n- Topics: function evaluation, domain/range, transformations (f(x)+k, f(x-h)), graphical analysis, composition f(g(x)), piecewise\n- At least one conceptual insight question (e.g., (p,q) and (q,p) on same graph)\n- Describe graphs in text. All computations yield clean results.\n\n[geometry_trigonometry] (MCQ + 1 SPR)\n- Topics: circle equations, Pythagorean theorem, special right triangles, coordinate geometry, trig ratios, volume/surface area, angle relationships\n- Optional \"visual\" field: {\"type\":\"coordinate_plane\",\"description\":\"...\"}\n- Radical answers OK (e.g., $5\\sqrt{13}$). Multi-step reasoning required.\n\n[statistics_probability] (MCQ + 1 SPR)\n- Topics: mean/median/mode, standard deviation (conceptual), probability (conditional from tables), data interpretation, scatter plots/line of best fit, sampling/bias\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Distractors: P(A|B) vs P(B|A) confusion, mean vs median, causal from correlational\n\n[advanced_math] (MCQ + 1 SPR)\n- Topics: quadratics (factoring, discriminant), polynomials (zeros, end behavior), exponentials (growth/decay), rational expressions, radical expressions, non-linear systems\n- Hardest questions on the test. Multi-step reasoning required. Elegant algebraic insight preferred.\n- All expressions must be factorable with integer coefficients. SPR answers are clean numbers.\n\nVERIFY EVERY ANSWER: Solve each problem step-by-step before finalizing correct_answer.\nOutput ONLY a valid JSON array of EXACTLY 5 objects. No markdown. Count them before outputting.", "text": "```json\n[\n {\n \"id\": \"sat99-math-m1-q18\",\n \"number\": 18,\n \"type\": \"statistics_probability\",\n \"prompt\": \"A scatterplot shows the relationship between the number of hours a student practices a musical instrument per week and the number of mistakes made during a performance. The line of best fit is given by the equation $y = -2.6x + 15$, where $y$ is the predicted number of mistakes and $x$ is the number of hours practiced. What is the best interpretation of the slope of the line of best fit?\",\n \"choices\": {\n \"A\": \"For each additional hour of practice, the predicted number of mistakes decreases by 2.6.\",\n \"B\": \"A student who does not practice is predicted to make 15 mistakes.\",\n \"C\": \"For every 2.6 additional mistakes made, the predicted number of practice hours decreases by 1.\",\n \"D\": \"The maximum possible number of mistakes is 15.\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"A\"\n },\n {\n \"id\": \"sat99-math-m1-q19\",\n \"number\": 19,\n \"type\": \"statistics_probability\",\n \"visual\": {\n \"type\": \"table\",\n \"title\": \"Participation in After-School Programs\",\n \"headers\": [\"\", \"Sports\", \"Arts\", \"Total\"],\n \"rows\": [\n [\"Juniors\", 85, 65, 150],\n [\"Seniors\", 95, 55, 150],\n [\"Total\", 180, 120, 300]\n ]\n },\n \"prompt\": \"The table above shows the distribution of 300 juniors and seniors at a high school based on their participation in after-school programs. If a student who participates in sports is selected at random, what is the probability that the student is a senior?\",\n \"choices\": {\n \"A\": \"$\\\\frac{95}{300}$\",\n \"B\": \"$\\\\frac{150}{300}$\",\n \"C\": \"$\\\\frac{95}{150}$\",\n \"D\": \"$\\\\frac{95}{180}$\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"D\"\n },\n {\n \"id\": \"sat99-math-m1-q20\",\n \"number\": 20,\n \"type\": \"statistics_probability\",\n \"prompt\": \"A data set consists of 9 integers. The numbers are 10, 12, 12, 13, 15, 17, 19, 20, and 45. Which of the following statements is true about the mean, median, and mode of this data set?\",\n \"choices\": {\n \"A\": \"The mean is greater than the median.\",\n \"B\": \"The median is greater than the mean.\",\n \"C\": \"The mode is greater than the mean.\",\n \"D\": \"The mean, median, and mode are all equal.\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"A\"\n },\n {\n \"id\": \"sat99-math-m1-q21\",\n \"number\": 21,\n \"type\": \"statistics_probability\",\n \"prompt\": \"An environmental scientist wants to estimate the proportion of trees in a large forest that are infected with a certain fungus. She divides the forest into 100 equally sized plots, randomly selects 10 plots, and examines every tree within those 10 plots. Which of the following best describes this sampling method?\",\n \"choices\": {\n \"A\": \"Simple random sampling\",\n \"B\": \"Cluster sampling\",\n \"C\": \"Stratified random sampling\",\n \"D\": \"Convenience sampling\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"B\"\n },\n {\n \"id\": \"sat99-math-m1-q22\",\n \"number\": 22,\n \"type\": \"statistics_probability\",\n \"visual\": {\n \"type\": \"table\",\n \"title\": \"Vehicle Types at a Service Center\",\n \"headers\": [\"\", \"Requires Oil Change\", \"Requires Tire Rotation\", \"Total\"],\n \"rows\": [\n [\"Sedan\", 50, 30, 80],\n [\"SUV\", 40, 35, 75],\n [\"Total\", 90, 65, 155]\n ]\n },\n \"prompt\": \"The table above shows the types of vehicles and the primary service requested at an auto service center on one day. Based on the table, what proportion of sedans required an oil change? (Express your answer as a decimal or fraction).\",\n \"answer_format\": \"student_produced\",\n \"correct_answer\": \"5/8\"\n }\n]\n```" }, { "costUsd": 0, "model": "gemini-2.5-pro", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 5 SAT Math questions (of which 2 are student-produced-response and the rest are MCQ).\nType: advanced_math | Test: 99 | Unit: 1 | Start Q#: 23\n\nMANDATORY: Your JSON array MUST contain exactly 5 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nMCQ Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"advanced_math\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\nSPR Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"advanced_math\",\"prompt\":\"...\",\"answer_format\":\"student_produced\",\"correct_answer\":\"numeric_value\"}\nPlace SPR questions at the END of the array.\n\nFollow instructions for your type:\n\n[algebra] (MCQ + 2 SPR)\n- Topics: linear equations (1 var, 2 vars), systems of 2 linear equations, linear inequalities, absolute value\n- Mix word problems and abstract algebra. Use $...$ notation.\n- MCQ distractors: sign errors, arithmetic mistakes, setup errors, partial answers\n- SPR answers: clean integers or simple fractions\n\n[functions_analysis] (MCQ + 2 SPR)\n- Topics: function evaluation, domain/range, transformations (f(x)+k, f(x-h)), graphical analysis, composition f(g(x)), piecewise\n- At least one conceptual insight question (e.g., (p,q) and (q,p) on same graph)\n- Describe graphs in text. All computations yield clean results.\n\n[geometry_trigonometry] (MCQ + 2 SPR)\n- Topics: circle equations, Pythagorean theorem, special right triangles, coordinate geometry, trig ratios, volume/surface area, angle relationships\n- Optional \"visual\" field: {\"type\":\"coordinate_plane\",\"description\":\"...\"}\n- Radical answers OK (e.g., $5\\sqrt{13}$). Multi-step reasoning required.\n\n[statistics_probability] (MCQ + 2 SPR)\n- Topics: mean/median/mode, standard deviation (conceptual), probability (conditional from tables), data interpretation, scatter plots/line of best fit, sampling/bias\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Distractors: P(A|B) vs P(B|A) confusion, mean vs median, causal from correlational\n\n[advanced_math] (MCQ + 2 SPR)\n- Topics: quadratics (factoring, discriminant), polynomials (zeros, end behavior), exponentials (growth/decay), rational expressions, radical expressions, non-linear systems\n- Hardest questions on the test. Multi-step reasoning required. Elegant algebraic insight preferred.\n- All expressions must be factorable with integer coefficients. SPR answers are clean numbers.\n\nVERIFY EVERY ANSWER: Solve each problem step-by-step before finalizing correct_answer.\nOutput ONLY a valid JSON array of EXACTLY 5 objects. No markdown. Count them before outputting.", "text": "```json\n[\n {\n \"id\": \"sat99-math-m1-q23\",\n \"number\": 23,\n \"type\": \"advanced_math\",\n \"prompt\": \"The quadratic equation $kx^2 - 6x + 3 = 0$, where $k$ is a constant, has exactly one real solution. What is the value of $k$?\",\n \"choices\": {\n \"A\": \"-3\",\n \"B\": \"0\",\n \"C\": \"3\",\n \"D\": \"9\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"C\"\n },\n {\n \"id\": \"sat99-math-m1-q24\",\n \"number\": 24,\n \"type\": \"advanced_math\",\n \"prompt\": \"The polynomial $P(x) = x^3 - 2x^2 - 5x + 6$ has a zero at $x=1$. What is the sum of the other two zeros of the polynomial?\",\n \"choices\": {\n \"A\": \"-1\",\n \"B\": \"1\",\n \"C\": \"2\",\n \"D\": \"5\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"B\"\n },\n {\n \"id\": \"sat99-math-m1-q25\",\n \"number\": 25,\n \"type\": \"advanced_math\",\n \"prompt\": \"If $x > 2$, which of the following expressions is equivalent to $\\\\frac{x - \\\\frac{4}{x}}{1 + \\\\frac{2}{x}}$?\",\n \"choices\": {\n \"A\": \"x - 2\",\n \"B\": \"x + 2\",\n \"C\": \"\\\\frac{x-4}{3}\",\n \"D\": \"x - 4\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"A\"\n },\n {\n \"id\": \"sat99-math-m1-q26\",\n \"number\": 26,\n \"type\": \"advanced_math\",\n \"prompt\": \"A system of two equations is given by $y = x^2 - 5x + 7$ and $y - x = 2$. If $(x, y)$ is a solution to the system, what is one possible value for $x$?\",\n \"answer_format\": \"student_produced\",\n \"correct_answer\": \"5\"\n },\n {\n \"id\": \"sat99-math-m1-q27\",\n \"number\": 27,\n \"type\": \"advanced_math\",\n \"prompt\": \"If $n > 0$ and $\\\\sqrt[3]{n^2} \\\\cdot n^k = n^{5/3}$, what is the value of $k$?\",\n \"answer_format\": \"student_produced\",\n \"correct_answer\": \"1\"\n }\n]\n```" }, { "costUsd": 0, "model": "gemini-2.5-pro", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 7 SAT Math questions (of which 2 are student-produced-response and the rest are MCQ).\nType: algebra | Test: 99 | Unit: 2 | Start Q#: 1\n\nMANDATORY: Your JSON array MUST contain exactly 7 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nMCQ Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"algebra\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\nSPR Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"algebra\",\"prompt\":\"...\",\"answer_format\":\"student_produced\",\"correct_answer\":\"numeric_value\"}\nPlace SPR questions at the END of the array.\n\nFollow instructions for your type:\n\n[algebra] (MCQ + 2 SPR)\n- Topics: linear equations (1 var, 2 vars), systems of 2 linear equations, linear inequalities, absolute value\n- Mix word problems and abstract algebra. Use $...$ notation.\n- MCQ distractors: sign errors, arithmetic mistakes, setup errors, partial answers\n- SPR answers: clean integers or simple fractions\n\n[functions_analysis] (MCQ + 2 SPR)\n- Topics: function evaluation, domain/range, transformations (f(x)+k, f(x-h)), graphical analysis, composition f(g(x)), piecewise\n- At least one conceptual insight question (e.g., (p,q) and (q,p) on same graph)\n- Describe graphs in text. All computations yield clean results.\n\n[geometry_trigonometry] (MCQ + 2 SPR)\n- Topics: circle equations, Pythagorean theorem, special right triangles, coordinate geometry, trig ratios, volume/surface area, angle relationships\n- Optional \"visual\" field: {\"type\":\"coordinate_plane\",\"description\":\"...\"}\n- Radical answers OK (e.g., $5\\sqrt{13}$). Multi-step reasoning required.\n\n[statistics_probability] (MCQ + 2 SPR)\n- Topics: mean/median/mode, standard deviation (conceptual), probability (conditional from tables), data interpretation, scatter plots/line of best fit, sampling/bias\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Distractors: P(A|B) vs P(B|A) confusion, mean vs median, causal from correlational\n\n[advanced_math] (MCQ + 2 SPR)\n- Topics: quadratics (factoring, discriminant), polynomials (zeros, end behavior), exponentials (growth/decay), rational expressions, radical expressions, non-linear systems\n- Hardest questions on the test. Multi-step reasoning required. Elegant algebraic insight preferred.\n- All expressions must be factorable with integer coefficients. SPR answers are clean numbers.\n\nVERIFY EVERY ANSWER: Solve each problem step-by-step before finalizing correct_answer.\nOutput ONLY a valid JSON array of EXACTLY 7 objects. No markdown. Count them before outputting.", "text": "```json\n[\n {\n \"id\": \"sat99-math-m2-q1\",\n \"number\": 1,\n \"type\": \"algebra\",\n \"prompt\": \"A car rental agency charges a flat daily fee of $40 plus an additional charge of $0.15 per mile driven. If a customer's total bill for a one-day rental was $98.50, how many miles did the customer drive?\",\n \"choices\": {\n \"A\": \"267\",\n \"B\": \"39\",\n \"C\": \"390\",\n \"D\": \"657\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"C\"\n },\n {\n \"id\": \"sat99-math-m2-q2\",\n \"number\": 2,\n \"type\": \"algebra\",\n \"prompt\": \"If $|2x - 3| = 11$, what is one possible value of $x$?\",\n \"choices\": {\n \"A\": \"7\",\n \"B\": \"4\",\n \"C\": \"-7\",\n \"D\": \"-4\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"D\"\n },\n {\n \"id\": \"sat99-math-m2-q3\",\n \"number\": 3,\n \"type\": \"algebra\",\n \"prompt\": \"At a food truck, 2 tacos and 1 burrito cost a total of $11.00. At the same food truck, 1 taco and 3 burritos cost a total of $18.00. What is the cost of one burrito?\",\n \"choices\": {\n \"A\": \"$2.50\",\n \"B\": \"$3.00\",\n \"C\": \"$5.50\",\n \"D\": \"$6.00\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"C\"\n },\n {\n \"id\": \"sat99-math-m2-q4\",\n \"number\": 4,\n \"type\": \"algebra\",\n \"prompt\": \"Which of the following values of $x$ is a solution to the inequality $8 - 3x < -7$?\",\n \"choices\": {\n \"A\": \"6\",\n \"B\": \"5\",\n \"C\": \"4\",\n \"D\": \"-5\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"A\"\n },\n {\n \"id\": \"sat99-math-m2-q5\",\n \"number\": 5,\n \"type\": \"algebra\",\n \"prompt\": \"A small business's weekly profit, $P$, in dollars, is modeled by the function $P(x) = 150x - 2700$, where $x$ represents the number of items sold. What is the minimum number of items the business must sell in a week to realize a profit?\",\n \"choices\": {\n \"A\": \"19\",\n \"B\": \"18\",\n \"C\": \"17\",\n \"D\": \"16\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"A\"\n },\n {\n \"id\": \"sat99-math-m2-q6\",\n \"number\": 6,\n \"type\": \"algebra\",\n \"prompt\": \"If $3x + 2y = 19$ and $x - y = 3$, what is the value of $x$?\",\n \"answer_format\": \"student_produced\",\n \"correct_answer\": \"5\"\n },\n {\n \"id\": \"sat99-math-m2-q7\",\n \"number\": 7,\n \"type\": \"algebra\",\n \"prompt\": \"If $\\\\frac{2}{3}(x - 4) = \\\\frac{1}{2}(x + 2)$, what is the value of $x$?\",\n \"answer_format\": \"student_produced\",\n \"correct_answer\": \"22\"\n }\n]\n```" }, { "costUsd": 0, "model": "gemini-2.5-pro", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 5 SAT Math questions (of which 1 are student-produced-response and the rest are MCQ).\nType: functions_analysis | Test: 99 | Unit: 2 | Start Q#: 8\n\nMANDATORY: Your JSON array MUST contain exactly 5 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nMCQ Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"functions_analysis\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\nSPR Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"functions_analysis\",\"prompt\":\"...\",\"answer_format\":\"student_produced\",\"correct_answer\":\"numeric_value\"}\nPlace SPR questions at the END of the array.\n\nFollow instructions for your type:\n\n[algebra] (MCQ + 1 SPR)\n- Topics: linear equations (1 var, 2 vars), systems of 2 linear equations, linear inequalities, absolute value\n- Mix word problems and abstract algebra. Use $...$ notation.\n- MCQ distractors: sign errors, arithmetic mistakes, setup errors, partial answers\n- SPR answers: clean integers or simple fractions\n\n[functions_analysis] (MCQ + 1 SPR)\n- Topics: function evaluation, domain/range, transformations (f(x)+k, f(x-h)), graphical analysis, composition f(g(x)), piecewise\n- At least one conceptual insight question (e.g., (p,q) and (q,p) on same graph)\n- Describe graphs in text. All computations yield clean results.\n\n[geometry_trigonometry] (MCQ + 1 SPR)\n- Topics: circle equations, Pythagorean theorem, special right triangles, coordinate geometry, trig ratios, volume/surface area, angle relationships\n- Optional \"visual\" field: {\"type\":\"coordinate_plane\",\"description\":\"...\"}\n- Radical answers OK (e.g., $5\\sqrt{13}$). Multi-step reasoning required.\n\n[statistics_probability] (MCQ + 1 SPR)\n- Topics: mean/median/mode, standard deviation (conceptual), probability (conditional from tables), data interpretation, scatter plots/line of best fit, sampling/bias\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Distractors: P(A|B) vs P(B|A) confusion, mean vs median, causal from correlational\n\n[advanced_math] (MCQ + 1 SPR)\n- Topics: quadratics (factoring, discriminant), polynomials (zeros, end behavior), exponentials (growth/decay), rational expressions, radical expressions, non-linear systems\n- Hardest questions on the test. Multi-step reasoning required. Elegant algebraic insight preferred.\n- All expressions must be factorable with integer coefficients. SPR answers are clean numbers.\n\nVERIFY EVERY ANSWER: Solve each problem step-by-step before finalizing correct_answer.\nOutput ONLY a valid JSON array of EXACTLY 5 objects. No markdown. Count them before outputting.", "text": "```json\n[\n {\n \"id\": \"sat99-math-m2-q8\",\n \"number\": 8,\n \"type\": \"functions_analysis\",\n \"prompt\": \"For the function $f$ defined by $f(x) = 2x^2 - 5x + 1$, what is the value of $f(3) - f(-1)$?\",\n \"choices\": {\n \"A\": \"-4\",\n \"B\": \"-1\",\n \"C\": \"4\",\n \"D\": \"12\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"A\"\n },\n {\n \"id\": \"sat99-math-m2-q9\",\n \"number\": 9,\n \"type\": \"functions_analysis\",\n \"prompt\": \"The graph of a function $g$ is shown in the $xy$-plane. The function $h$ is defined by $h(x) = g(x-2) + 3$. If the point $(-1, 4)$ is on the graph of $g$, what is a point on the graph of $h$?\",\n \"choices\": {\n \"A\": \"(-3, 7)\",\n \"B\": \"(-3, 1)\",\n \"C\": \"(1, 7)\",\n \"D\": \"(1, 1)\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"C\"\n },\n {\n \"id\": \"sat99-math-m2-q10\",\n \"number\": 10,\n \"type\": \"functions_analysis\",\n \"prompt\": \"The function $f$ is an even function, meaning $f(x) = f(-x)$ for all values of $x$. The function $g$ is an odd function, meaning $g(-x) = -g(x)$ for all values of $x$. If $f(2) = 5$ and $g(-3) = -4$, what is the value of $f(-2) + g(3)$?\",\n \"choices\": {\n \"A\": \"-9\",\n \"B\": \"1\",\n \"C\": \"9\",\n \"D\": \"The value cannot be determined from the information given.\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"C\"\n },\n {\n \"id\": \"sat99-math-m2-q11\",\n \"number\": 11,\n \"type\": \"functions_analysis\",\n \"prompt\": \"Let the functions $p$ and $q$ be defined by $p(x) = x^2 - 4$ and $q(x) = 2x + 1$. For what positive value of $x$ is $p(q(x)) = 0$?\",\n \"choices\": {\n \"A\": \"$\\\\frac{1}{2}$\",\n \"B\": \"$\\\\frac{3}{2}$\",\n \"C\": \"$2$\",\n \"D\": \"$5$\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"A\"\n },\n {\n \"id\": \"sat99-math-m2-q12\",\n \"number\": 12,\n \"type\": \"functions_analysis\",\n \"prompt\": \"The function $f$ is defined below.\\n$f(x) = \\\\begin{cases} x^2+3 & \\\\text{if } x < 1 \\\\\\\\ 2x-1 & \\\\text{if } x \\\\ge 1 \\\\end{cases}$\\nWhat is the value of $f(f(0))$?\",\n \"answer_format\": \"student_produced\",\n \"correct_answer\": \"5\"\n }\n]\n```" }, { "costUsd": 0, "model": "gemini-2.5-pro", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 5 SAT Math questions (of which 1 are student-produced-response and the rest are MCQ).\nType: geometry_trigonometry | Test: 99 | Unit: 2 | Start Q#: 13\n\nMANDATORY: Your JSON array MUST contain exactly 5 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nMCQ Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"geometry_trigonometry\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\nSPR Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"geometry_trigonometry\",\"prompt\":\"...\",\"answer_format\":\"student_produced\",\"correct_answer\":\"numeric_value\"}\nPlace SPR questions at the END of the array.\n\nFollow instructions for your type:\n\n[algebra] (MCQ + 1 SPR)\n- Topics: linear equations (1 var, 2 vars), systems of 2 linear equations, linear inequalities, absolute value\n- Mix word problems and abstract algebra. Use $...$ notation.\n- MCQ distractors: sign errors, arithmetic mistakes, setup errors, partial answers\n- SPR answers: clean integers or simple fractions\n\n[functions_analysis] (MCQ + 1 SPR)\n- Topics: function evaluation, domain/range, transformations (f(x)+k, f(x-h)), graphical analysis, composition f(g(x)), piecewise\n- At least one conceptual insight question (e.g., (p,q) and (q,p) on same graph)\n- Describe graphs in text. All computations yield clean results.\n\n[geometry_trigonometry] (MCQ + 1 SPR)\n- Topics: circle equations, Pythagorean theorem, special right triangles, coordinate geometry, trig ratios, volume/surface area, angle relationships\n- Optional \"visual\" field: {\"type\":\"coordinate_plane\",\"description\":\"...\"}\n- Radical answers OK (e.g., $5\\sqrt{13}$). Multi-step reasoning required.\n\n[statistics_probability] (MCQ + 1 SPR)\n- Topics: mean/median/mode, standard deviation (conceptual), probability (conditional from tables), data interpretation, scatter plots/line of best fit, sampling/bias\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Distractors: P(A|B) vs P(B|A) confusion, mean vs median, causal from correlational\n\n[advanced_math] (MCQ + 1 SPR)\n- Topics: quadratics (factoring, discriminant), polynomials (zeros, end behavior), exponentials (growth/decay), rational expressions, radical expressions, non-linear systems\n- Hardest questions on the test. Multi-step reasoning required. Elegant algebraic insight preferred.\n- All expressions must be factorable with integer coefficients. SPR answers are clean numbers.\n\nVERIFY EVERY ANSWER: Solve each problem step-by-step before finalizing correct_answer.\nOutput ONLY a valid JSON array of EXACTLY 5 objects. No markdown. Count them before outputting.", "text": "```json\n[\n {\n \"id\": \"sat99-math-m2-q13\",\n \"number\": 13,\n \"type\": \"geometry_trigonometry\",\n \"prompt\": \"A circle in the $xy$-plane has its center at $(3, -5)$. The point $(8, 7)$ lies on the circle. Which of the following is the equation of the circle?\",\n \"choices\": {\n \"A\": \"$(x-3)^2 + (y+5)^2 = 13$\",\n \"B\": \"$(x-8)^2 + (y-7)^2 = 169$\",\n \"C\": \"$(x+3)^2 + (y-5)^2 = 169$\",\n \"D\": \"$(x-3)^2 + (y+5)^2 = 169$\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"D\"\n },\n {\n \"id\": \"sat99-math-m2-q14\",\n \"number\": 14,\n \"type\": \"geometry_trigonometry\",\n \"prompt\": \"In right triangle $ABC$, the measure of angle $C$ is $90^{\\\\circ}$. If $\\\\sin(A) = \\\\frac{8}{17}$, what is the value of $\\\\cos(B)$?\",\n \"choices\": {\n \"A\": \"$\\\\frac{15}{17}$\",\n \"B\": \"$\\\\frac{8}{15}$\",\n \"C\": \"$\\\\frac{8}{17}$\",\n \"D\": \"$\\\\frac{17}{8}$\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"C\"\n },\n {\n \"id\": \"sat99-math-m2-q15\",\n \"number\": 15,\n \"type\": \"geometry_trigonometry\",\n \"prompt\": \"A cylindrical grain silo has a radius of 10 feet and a height of 40 feet. The top of the silo is a hemisphere with the same radius. What is the total volume of the silo in cubic feet?\",\n \"choices\": {\n \"A\": \"$4000\\\\pi$\",\n \"B\": \"$\\\\frac{14000}{3}\\\\pi$\",\n \"C\": \"$\\\\frac{16000}{3}\\\\pi$\",\n \"D\": \"$4400\\\\pi$\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"B\"\n },\n {\n \"id\": \"sat99-math-m2-q16\",\n \"number\": 16,\n \"type\": \"geometry_trigonometry\",\n \"prompt\": \"In the $xy$-plane, line $L$ is defined by the equation $y = -\\\\frac{1}{3}x + 5$. Line $M$ is perpendicular to line $L$ and passes through the point $(-2, 2)$. What is the $y$-intercept of line $M$?\",\n \"choices\": {\n \"A\": \"8\",\n \"B\": \"-8\",\n \"C\": \"3\",\n \"D\": \"$-\\\\frac{1}{3}$\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"A\"\n },\n {\n \"id\": \"sat99-math-m2-q17\",\n \"number\": 17,\n \"type\": \"geometry_trigonometry\",\n \"prompt\": \"A ladder is placed against a vertical wall. The base of the ladder is 7 feet away from the base of the wall, and the ladder makes an angle of $60^{\\\\circ}$ with the ground. What is the length of the ladder in feet?\",\n \"answer_format\": \"student_produced\",\n \"correct_answer\": \"14\"\n }\n]\n```" }, { "costUsd": 0, "model": "gemini-2.5-pro", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 5 SAT Math questions (of which 1 are student-produced-response and the rest are MCQ).\nType: statistics_probability | Test: 99 | Unit: 2 | Start Q#: 18\n\nMANDATORY: Your JSON array MUST contain exactly 5 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nMCQ Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"statistics_probability\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\nSPR Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"statistics_probability\",\"prompt\":\"...\",\"answer_format\":\"student_produced\",\"correct_answer\":\"numeric_value\"}\nPlace SPR questions at the END of the array.\n\nFollow instructions for your type:\n\n[algebra] (MCQ + 1 SPR)\n- Topics: linear equations (1 var, 2 vars), systems of 2 linear equations, linear inequalities, absolute value\n- Mix word problems and abstract algebra. Use $...$ notation.\n- MCQ distractors: sign errors, arithmetic mistakes, setup errors, partial answers\n- SPR answers: clean integers or simple fractions\n\n[functions_analysis] (MCQ + 1 SPR)\n- Topics: function evaluation, domain/range, transformations (f(x)+k, f(x-h)), graphical analysis, composition f(g(x)), piecewise\n- At least one conceptual insight question (e.g., (p,q) and (q,p) on same graph)\n- Describe graphs in text. All computations yield clean results.\n\n[geometry_trigonometry] (MCQ + 1 SPR)\n- Topics: circle equations, Pythagorean theorem, special right triangles, coordinate geometry, trig ratios, volume/surface area, angle relationships\n- Optional \"visual\" field: {\"type\":\"coordinate_plane\",\"description\":\"...\"}\n- Radical answers OK (e.g., $5\\sqrt{13}$). Multi-step reasoning required.\n\n[statistics_probability] (MCQ + 1 SPR)\n- Topics: mean/median/mode, standard deviation (conceptual), probability (conditional from tables), data interpretation, scatter plots/line of best fit, sampling/bias\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Distractors: P(A|B) vs P(B|A) confusion, mean vs median, causal from correlational\n\n[advanced_math] (MCQ + 1 SPR)\n- Topics: quadratics (factoring, discriminant), polynomials (zeros, end behavior), exponentials (growth/decay), rational expressions, radical expressions, non-linear systems\n- Hardest questions on the test. Multi-step reasoning required. Elegant algebraic insight preferred.\n- All expressions must be factorable with integer coefficients. SPR answers are clean numbers.\n\nVERIFY EVERY ANSWER: Solve each problem step-by-step before finalizing correct_answer.\nOutput ONLY a valid JSON array of EXACTLY 5 objects. No markdown. Count them before outputting.", "text": "```json\n[\n {\n \"id\": \"sat99-math-m2-q18\",\n \"number\": 18,\n \"type\": \"statistics_probability\",\n \"prompt\": \"A market research firm surveyed 200 owners of either a sedan or an SUV. Each owner was asked whether their next vehicle purchase would be a new or used vehicle. The results are partially summarized in the table below. Of those surveyed, 60% owned a sedan.\",\n \"visual\": {\n \"type\": \"table\",\n \"title\": \"Next Vehicle Purchase Plan\",\n \"headers\": [\"Current Vehicle\", \"New\", \"Used\", \"Total\"],\n \"rows\": [\n [\"Sedan\", 72, \"\", \"\"],\n [\"SUV\", \"\", \"\", \"\"],\n [\"Total\", 110, 90, 200]\n ]\n },\n \"prompt_continue\": \"Based on the table, what is the probability that a person who currently owns an SUV will plan to purchase a new vehicle?\",\n \"choices\": {\n \"A\": \"0.34\",\n \"B\": \"0.475\",\n \"C\": \"0.525\",\n \"D\": \"0.65\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"B\"\n },\n {\n \"id\": \"sat99-math-m2-q19\",\n \"number\": 19,\n \"type\": \"statistics_probability\",\n \"prompt\": \"The dot plot above shows the number of goals scored by a soccer team in 15 games. If the game with the highest number of goals is removed from the data, which of the following statements about the mean and median of the resulting data is true?\",\n \"visual\": {\n \"type\": \"dot_plot\",\n \"description\": \"A dot plot titled 'Goals Scored in 15 Games'. The x-axis is labeled 'Number of Goals'. There are 3 dots at 0, 5 dots at 1, 4 dots at 2, 2 dots at 3, and 1 dot at 6.\"\n },\n \"choices\": {\n \"A\": \"The mean will decrease, and the median will remain the same.\",\n \"B\": \"The mean will decrease, and the median will decrease.\",\n \"C\": \"The mean will remain the same, and the median will decrease.\",\n \"D\": \"Both the mean and the median will remain the same.\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"A\"\n },\n {\n \"id\": \"sat99-math-m2-q20\",\n \"number\": 20,\n \"type\": \"statistics_probability\",\n \"prompt\": \"A sample of 50 trees was selected from a large forest. The height, in feet, and the trunk diameter, in inches, was measured for each tree. The data are shown in the scatterplot, along with a line of best fit. The equation of the line is $h = 3.8d + 15.2$, where $h$ is the height and $d$ is the trunk diameter.\",\n \"visual\": {\n \"type\": \"scatter_plot\",\n \"description\": \"A scatterplot titled 'Tree Height vs. Trunk Diameter' showing a positive linear association. The x-axis is 'Trunk Diameter (inches)' from 5 to 25. The y-axis is 'Height (feet)' from 20 to 120. A line of best fit with a positive slope is drawn through the data points.\"\n },\n \"prompt_continue\": \"Which of the following is the best interpretation of the slope of the line of best fit?\",\n \"choices\": {\n \"A\": \"For a tree with a trunk diameter of 0 inches, the predicted height is 3.8 feet.\",\n \"B\": \"For every 1-foot increase in a tree's height, the predicted trunk diameter increases by 3.8 inches.\",\n \"C\": \"For every 3.8-inch increase in a tree's trunk diameter, the predicted height increases by 1 foot.\",\n \"D\": \"For every 1-inch increase in a tree's trunk diameter, the predicted height increases by 3.8 feet.\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"D\"\n },\n {\n \"id\": \"sat99-math-m2-q21\",\n \"number\": 21,\n \"type\": \"statistics_probability\",\n \"prompt\": \"The test scores for two history classes are summarized by the box plots below.\",\n \"visual\": {\n \"type\": \"box_plot\",\n \"description\": \"Two box plots on a single axis from 50 to 100. 'Class A': min=55, Q1=65, median=75, Q3=80, max=90. 'Class B': min=60, Q1=75, median=80, Q3=85, max=90. The box for Class A is wider than the box for Class B, indicating a larger interquartile range.\"\n },\n \"prompt_continue\": \"Based on the box plots, which of the following must be true?\",\n \"choices\": {\n \"A\": \"The mean score of Class A is greater than the mean score of Class B.\",\n \"B\": \"The range of scores for Class A is greater than the range of scores for Class B.\",\n \"C\": \"The median score of Class B is greater than the median score of Class A.\",\n \"D\": \"There are more students in Class B than in Class A.\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"C\"\n },\n {\n \"id\": \"sat99-math-m2-q22\",\n \"number\": 22,\n \"type\": \"statistics_probability\",\n \"prompt\": \"A dataset consists of 7 distinct integers. The median of the dataset is 12, the minimum value is 5, and the range is 14. What is the maximum possible value of the mean of the dataset, rounded to the nearest tenth?\",\n \"answer_format\": \"student_produced\",\n \"correct_answer\": \"14.3\"\n }\n]\n```" }, { "costUsd": 0, "model": "gemini-2.5-pro", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 5 SAT Math questions (of which 2 are student-produced-response and the rest are MCQ).\nType: advanced_math | Test: 99 | Unit: 2 | Start Q#: 23\n\nMANDATORY: Your JSON array MUST contain exactly 5 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nMCQ Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"advanced_math\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\nSPR Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"advanced_math\",\"prompt\":\"...\",\"answer_format\":\"student_produced\",\"correct_answer\":\"numeric_value\"}\nPlace SPR questions at the END of the array.\n\nFollow instructions for your type:\n\n[algebra] (MCQ + 2 SPR)\n- Topics: linear equations (1 var, 2 vars), systems of 2 linear equations, linear inequalities, absolute value\n- Mix word problems and abstract algebra. Use $...$ notation.\n- MCQ distractors: sign errors, arithmetic mistakes, setup errors, partial answers\n- SPR answers: clean integers or simple fractions\n\n[functions_analysis] (MCQ + 2 SPR)\n- Topics: function evaluation, domain/range, transformations (f(x)+k, f(x-h)), graphical analysis, composition f(g(x)), piecewise\n- At least one conceptual insight question (e.g., (p,q) and (q,p) on same graph)\n- Describe graphs in text. All computations yield clean results.\n\n[geometry_trigonometry] (MCQ + 2 SPR)\n- Topics: circle equations, Pythagorean theorem, special right triangles, coordinate geometry, trig ratios, volume/surface area, angle relationships\n- Optional \"visual\" field: {\"type\":\"coordinate_plane\",\"description\":\"...\"}\n- Radical answers OK (e.g., $5\\sqrt{13}$). Multi-step reasoning required.\n\n[statistics_probability] (MCQ + 2 SPR)\n- Topics: mean/median/mode, standard deviation (conceptual), probability (conditional from tables), data interpretation, scatter plots/line of best fit, sampling/bias\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Distractors: P(A|B) vs P(B|A) confusion, mean vs median, causal from correlational\n\n[advanced_math] (MCQ + 2 SPR)\n- Topics: quadratics (factoring, discriminant), polynomials (zeros, end behavior), exponentials (growth/decay), rational expressions, radical expressions, non-linear systems\n- Hardest questions on the test. Multi-step reasoning required. Elegant algebraic insight preferred.\n- All expressions must be factorable with integer coefficients. SPR answers are clean numbers.\n\nVERIFY EVERY ANSWER: Solve each problem step-by-step before finalizing correct_answer.\nOutput ONLY a valid JSON array of EXACTLY 5 objects. No markdown. Count them before outputting.", "text": "```json\n[\n {\n \"id\": \"sat99-math-m2-q23\",\n \"number\": 23,\n \"type\": \"advanced_math\",\n \"prompt\": \"The system of equations $y = x^2 - 4x + 7$ and $y = 2x + 2$ is graphed in the $xy$-plane. What is the sum of the $y$-coordinates of the two intersection points?\",\n \"choices\": {\n \"A\": \"6\",\n \"B\": \"12\",\n \"C\": \"18\",\n \"D\": \"24\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"C\"\n },\n {\n \"id\": \"sat99-math-m2-q24\",\n \"number\": 24,\n \"type\": \"advanced_math\",\n \"prompt\": \"A polynomial function $p(x)$ has zeros at $x=-1$ and $x=4$, where the zero at $x=4$ has a multiplicity of 2. The graph of $y=p(x)$ has a positive $y$-intercept. Which of the following could be the equation for $p(x)$?\",\n \"choices\": {\n \"A\": \"p(x) = (x+1)(x-4)^2\",\n \"B\": \"p(x) = (x-1)(x+4)^2\",\n \"C\": \"p(x) = -(x-1)(x+4)^2\",\n \"D\": \"p(x) = -(x+1)(x-4)^2\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"D\"\n },\n {\n \"id\": \"sat99-math-m2-q25\",\n \"number\": 25,\n \"type\": \"advanced_math\",\n \"prompt\": \"The function $f$ is defined by $f(x) = \\\\frac{2x^2 + 7x - 15}{x+5}$. The graph of $y=f(x)$ is a line with a point of discontinuity. What is the $y$-coordinate of the point of discontinuity?\",\n \"choices\": {\n \"A\": \"-13\",\n \"B\": \"-5\",\n \"C\": \"2\",\n \"D\": \"3\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"A\"\n },\n {\n \"id\": \"sat99-math-m2-q26\",\n \"number\": 26,\n \"type\": \"advanced_math\",\n \"prompt\": \"If $x > 0$, what value of $x$ is the solution to the equation $\\\\sqrt{2x+3} = x$?\",\n \"answer_format\": \"student_produced\",\n \"correct_answer\": \"3\"\n },\n {\n \"id\": \"sat99-math-m2-q27\",\n \"number\": 27,\n \"type\": \"advanced_math\",\n \"prompt\": \"The value of an investment is modeled by the function $V(t) = 100(1.5)^{\\\\frac{t}{4}}$, where $V(t)$ is the value of the investment in thousands of dollars and $t$ is the number of years since the initial investment was made. After how many years will the value of the investment be $225,000?\",\n \"answer_format\": \"student_produced\",\n \"correct_answer\": \"8\"\n }\n]\n```" } ], "results": [ { "attempts": 1, "index": 0, "output": { "costUsd": 0, "model": "gemini-2.5-pro", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 7 SAT Math questions (of which 2 are student-produced-response and the rest are MCQ).\nType: algebra | Test: 99 | Unit: 1 | Start Q#: 1\n\nMANDATORY: Your JSON array MUST contain exactly 7 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nMCQ Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"algebra\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\nSPR Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"algebra\",\"prompt\":\"...\",\"answer_format\":\"student_produced\",\"correct_answer\":\"numeric_value\"}\nPlace SPR questions at the END of the array.\n\nFollow instructions for your type:\n\n[algebra] (MCQ + 2 SPR)\n- Topics: linear equations (1 var, 2 vars), systems of 2 linear equations, linear inequalities, absolute value\n- Mix word problems and abstract algebra. Use $...$ notation.\n- MCQ distractors: sign errors, arithmetic mistakes, setup errors, partial answers\n- SPR answers: clean integers or simple fractions\n\n[functions_analysis] (MCQ + 2 SPR)\n- Topics: function evaluation, domain/range, transformations (f(x)+k, f(x-h)), graphical analysis, composition f(g(x)), piecewise\n- At least one conceptual insight question (e.g., (p,q) and (q,p) on same graph)\n- Describe graphs in text. All computations yield clean results.\n\n[geometry_trigonometry] (MCQ + 2 SPR)\n- Topics: circle equations, Pythagorean theorem, special right triangles, coordinate geometry, trig ratios, volume/surface area, angle relationships\n- Optional \"visual\" field: {\"type\":\"coordinate_plane\",\"description\":\"...\"}\n- Radical answers OK (e.g., $5\\sqrt{13}$). Multi-step reasoning required.\n\n[statistics_probability] (MCQ + 2 SPR)\n- Topics: mean/median/mode, standard deviation (conceptual), probability (conditional from tables), data interpretation, scatter plots/line of best fit, sampling/bias\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Distractors: P(A|B) vs P(B|A) confusion, mean vs median, causal from correlational\n\n[advanced_math] (MCQ + 2 SPR)\n- Topics: quadratics (factoring, discriminant), polynomials (zeros, end behavior), exponentials (growth/decay), rational expressions, radical expressions, non-linear systems\n- Hardest questions on the test. Multi-step reasoning required. Elegant algebraic insight preferred.\n- All expressions must be factorable with integer coefficients. SPR answers are clean numbers.\n\nVERIFY EVERY ANSWER: Solve each problem step-by-step before finalizing correct_answer.\nOutput ONLY a valid JSON array of EXACTLY 7 objects. No markdown. Count them before outputting.", "text": "```json\n[\n {\n \"id\": \"sat99-math-m1-q1\",\n \"number\": 1,\n \"type\": \"algebra\",\n \"prompt\": \"A moving company charges a flat fee of $120 for using their truck and an additional $45 per hour for labor. If a customer's final bill was $390, how many hours of labor did the customer pay for?\",\n \"choices\": {\n \"A\": \"2.6\",\n \"B\": \"5\",\n \"C\": \"6\",\n \"D\": \"8.6\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"C\"\n },\n {\n \"id\": \"sat99-math-m1-q2\",\n \"number\": 2,\n \"type\": \"algebra\",\n \"prompt\": \"At a school fundraiser, adults' tickets were sold for $8 and students' tickets were sold for $5. A total of 150 tickets were sold, bringing in a total of $990. How many adult tickets were sold?\",\n \"choices\": {\n \"A\": \"70\",\n \"B\": \"80\",\n \"C\": \"90\",\n \"D\": \"100\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"B\"\n },\n {\n \"id\": \"sat99-math-m1-q3\",\n \"number\": 3,\n \"type\": \"algebra\",\n \"prompt\": \"The atmospheric temperature, $T$, in degrees Celsius, required for a certain scientific experiment must satisfy the inequality $|T - (-20)| \\\\le 1.5$. Which of the following represents all possible values for the temperature $T$?\",\n \"choices\": {\n \"A\": \"T \\\\ge -18.5\",\n \"B\": \"T \\\\le -21.5\",\n \"C\": \"-18.5 \\\\le T \\\\le 18.5\",\n \"D\": \"-21.5 \\\\le T \\\\le -18.5\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"D\"\n },\n {\n \"id\": \"sat99-math-m1-q4\",\n \"number\": 4,\n \"type\": \"algebra\",\n \"prompt\": \"The total cost, $C$, in dollars, for an online subscription is given by the equation $C = 15m + 20$, where $m$ is the number of months. What is the best interpretation of the number 15 in this context?\",\n \"choices\": {\n \"A\": \"The cost per month\",\n \"B\": \"The initial sign-up fee\",\n \"C\": \"The total cost for 20 months\",\n \"D\": \"The number of months included in the subscription\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"A\"\n },\n {\n \"id\": \"sat99-math-m1-q5\",\n \"number\": 5,\n \"type\": \"algebra\",\n \"prompt\": \"If $2x - 3y = 8$ and $x + y = 9$, what is the value of the expression $x - 4y$?\",\n \"choices\": {\n \"A\": \"17\",\n \"B\": \"-1\",\n \"C\": \"1\",\n \"D\": \"7\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"B\"\n },\n {\n \"id\": \"sat99-math-m1-q6\",\n \"number\": 6,\n \"type\": \"algebra\",\n \"prompt\": \"For a certain chemical process, the concentration $C$ of a substance must be maintained such that $|3C - 11| = 4$. What is one possible value for the concentration $C$?\",\n \"answer_format\": \"student_produced\",\n \"correct_answer\": \"5\"\n },\n {\n \"id\": \"sat99-math-m1-q7\",\n \"number\": 7,\n \"type\": \"algebra\",\n \"prompt\": \"If $x > 0$ and $5x - 2y = 15$ and $4x - y = 12$, what is the value of $x$?\",\n \"answer_format\": \"student_produced\",\n \"correct_answer\": \"3\"\n }\n]\n```" }, "status": "ok" }, { "attempts": 1, "index": 1, "output": { "costUsd": 0, "model": "gemini-2.5-pro", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 5 SAT Math questions (of which 1 are student-produced-response and the rest are MCQ).\nType: functions_analysis | Test: 99 | Unit: 1 | Start Q#: 8\n\nMANDATORY: Your JSON array MUST contain exactly 5 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nMCQ Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"functions_analysis\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\nSPR Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"functions_analysis\",\"prompt\":\"...\",\"answer_format\":\"student_produced\",\"correct_answer\":\"numeric_value\"}\nPlace SPR questions at the END of the array.\n\nFollow instructions for your type:\n\n[algebra] (MCQ + 1 SPR)\n- Topics: linear equations (1 var, 2 vars), systems of 2 linear equations, linear inequalities, absolute value\n- Mix word problems and abstract algebra. Use $...$ notation.\n- MCQ distractors: sign errors, arithmetic mistakes, setup errors, partial answers\n- SPR answers: clean integers or simple fractions\n\n[functions_analysis] (MCQ + 1 SPR)\n- Topics: function evaluation, domain/range, transformations (f(x)+k, f(x-h)), graphical analysis, composition f(g(x)), piecewise\n- At least one conceptual insight question (e.g., (p,q) and (q,p) on same graph)\n- Describe graphs in text. All computations yield clean results.\n\n[geometry_trigonometry] (MCQ + 1 SPR)\n- Topics: circle equations, Pythagorean theorem, special right triangles, coordinate geometry, trig ratios, volume/surface area, angle relationships\n- Optional \"visual\" field: {\"type\":\"coordinate_plane\",\"description\":\"...\"}\n- Radical answers OK (e.g., $5\\sqrt{13}$). Multi-step reasoning required.\n\n[statistics_probability] (MCQ + 1 SPR)\n- Topics: mean/median/mode, standard deviation (conceptual), probability (conditional from tables), data interpretation, scatter plots/line of best fit, sampling/bias\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Distractors: P(A|B) vs P(B|A) confusion, mean vs median, causal from correlational\n\n[advanced_math] (MCQ + 1 SPR)\n- Topics: quadratics (factoring, discriminant), polynomials (zeros, end behavior), exponentials (growth/decay), rational expressions, radical expressions, non-linear systems\n- Hardest questions on the test. Multi-step reasoning required. Elegant algebraic insight preferred.\n- All expressions must be factorable with integer coefficients. SPR answers are clean numbers.\n\nVERIFY EVERY ANSWER: Solve each problem step-by-step before finalizing correct_answer.\nOutput ONLY a valid JSON array of EXACTLY 5 objects. No markdown. Count them before outputting.", "text": "```json\n[\n {\n \"id\": \"sat99-math-m1-q8\",\n \"number\": 8,\n \"type\": \"functions_analysis\",\n \"prompt\": \"Let the function $f$ be defined as $f(x) = x^2 + 5$ and the function $g$ be defined as $g(x) = x - 2$. What is the value of $f(g(3))$?\",\n \"choices\": {\n \"A\": \"6\",\n \"B\": \"8\",\n \"C\": \"14\",\n \"D\": \"17\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"A\"\n },\n {\n \"id\": \"sat99-math-m1-q9\",\n \"number\": 9,\n \"type\": \"functions_analysis\",\n \"prompt\": \"The graph of $y = h(x)$ is shown in the $xy$-plane. A new function, $k(x)$, is defined as $k(x) = h(x) - 2$. What is the maximum value of the function $k(x)$?\",\n \"visual\": {\n \"type\": \"coordinate_plane\",\n \"description\": \"A smooth curve representing the function h(x). The curve starts at (-4, -1), goes up to a maximum point at (-2, 5), goes down to a minimum point at (1, 0), and then goes up to the point (3, 3).\"\n },\n \"choices\": {\n \"A\": \"-2\",\n \"B\": \"3\",\n \"C\": \"5\",\n \"D\": \"7\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"B\"\n },\n {\n \"id\": \"sat99-math-m1-q10\",\n \"number\": 10,\n \"type\": \"functions_analysis\",\n \"prompt\": \"A function $f(x)$ has a domain of $x \\ge 2$ and a range of $y \\le 5$. The function $g(x)$ is defined as $g(x) = f(x+3) - 4$. Which of the following represents the domain and range of $g(x)$?\",\n \"choices\": {\n \"A\": \"Domain: $x \\ge 5$, Range: $y \\le 1$\",\n \"B\": \"Domain: $x \\ge -1$, Range: $y \\le 9$\",\n \"C\": \"Domain: $x \\ge 5$, Range: $y \\le 9$\",\n \"D\": \"Domain: $x \\ge -1$, Range: $y \\le 1$\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"D\"\n },\n {\n \"id\": \"sat99-math-m1-q11\",\n \"number\": 11,\n \"type\": \"functions_analysis\",\n \"prompt\": \"For a certain function $f$, it is known that if the point $(a, b)$ is on the graph of $y=f(x)$, then the point $(b, a)$ is also on the graph. If $f(2) = 5$, what is the value of $f(f(2))$?\",\n \"choices\": {\n \"A\": \"0\",\n \"B\": \"2\",\n \"C\": \"5\",\n \"D\": \"The value cannot be determined from the information given.\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"B\"\n },\n {\n \"id\": \"sat99-math-m1-q12\",\n \"number\": 12,\n \"type\": \"functions_analysis\",\n \"prompt\": \"For the piecewise function $g(x)$ defined below,\\n$g(x) = \\\\begin{cases} 4x-3 & \\\\text{if } x < 2 \\\\\\\\ x^2+1 & \\\\text{if } x \\\\ge 2 \\\\end{cases}$\\nWhat is the value of $g(g(1))$?\",\n \"answer_format\": \"student_produced\",\n \"correct_answer\": \"2\"\n }\n]\n```" }, "status": "ok" }, { "attempts": 1, "index": 2, "output": { "costUsd": 0, "model": "gemini-2.5-pro", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 5 SAT Math questions (of which 1 are student-produced-response and the rest are MCQ).\nType: geometry_trigonometry | Test: 99 | Unit: 1 | Start Q#: 13\n\nMANDATORY: Your JSON array MUST contain exactly 5 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nMCQ Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"geometry_trigonometry\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\nSPR Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"geometry_trigonometry\",\"prompt\":\"...\",\"answer_format\":\"student_produced\",\"correct_answer\":\"numeric_value\"}\nPlace SPR questions at the END of the array.\n\nFollow instructions for your type:\n\n[algebra] (MCQ + 1 SPR)\n- Topics: linear equations (1 var, 2 vars), systems of 2 linear equations, linear inequalities, absolute value\n- Mix word problems and abstract algebra. Use $...$ notation.\n- MCQ distractors: sign errors, arithmetic mistakes, setup errors, partial answers\n- SPR answers: clean integers or simple fractions\n\n[functions_analysis] (MCQ + 1 SPR)\n- Topics: function evaluation, domain/range, transformations (f(x)+k, f(x-h)), graphical analysis, composition f(g(x)), piecewise\n- At least one conceptual insight question (e.g., (p,q) and (q,p) on same graph)\n- Describe graphs in text. All computations yield clean results.\n\n[geometry_trigonometry] (MCQ + 1 SPR)\n- Topics: circle equations, Pythagorean theorem, special right triangles, coordinate geometry, trig ratios, volume/surface area, angle relationships\n- Optional \"visual\" field: {\"type\":\"coordinate_plane\",\"description\":\"...\"}\n- Radical answers OK (e.g., $5\\sqrt{13}$). Multi-step reasoning required.\n\n[statistics_probability] (MCQ + 1 SPR)\n- Topics: mean/median/mode, standard deviation (conceptual), probability (conditional from tables), data interpretation, scatter plots/line of best fit, sampling/bias\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Distractors: P(A|B) vs P(B|A) confusion, mean vs median, causal from correlational\n\n[advanced_math] (MCQ + 1 SPR)\n- Topics: quadratics (factoring, discriminant), polynomials (zeros, end behavior), exponentials (growth/decay), rational expressions, radical expressions, non-linear systems\n- Hardest questions on the test. Multi-step reasoning required. Elegant algebraic insight preferred.\n- All expressions must be factorable with integer coefficients. SPR answers are clean numbers.\n\nVERIFY EVERY ANSWER: Solve each problem step-by-step before finalizing correct_answer.\nOutput ONLY a valid JSON array of EXACTLY 5 objects. No markdown. Count them before outputting.", "text": "```json\n[\n {\n \"id\": \"sat99-math-m1-q13\",\n \"number\": 13,\n \"type\": \"geometry_trigonometry\",\n \"prompt\": \"The equation of a circle in the $xy$-plane is $x^2 + y^2 - 6x + 8y + 21 = 0$. What is the radius of the circle?\",\n \"choices\": {\n \"A\": \"5\",\n \"B\": \"$\\\\sqrt{21}$\",\n \"C\": \"2\",\n \"D\": \"4\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"C\"\n },\n {\n \"id\": \"sat99-math-m1-q14\",\n \"number\": 14,\n \"type\": \"geometry_trigonometry\",\n \"prompt\": \"In the $xy$-plane, a line segment is drawn from the origin to a point $P$ in Quadrant I. If the length of the segment is 8 and it makes an angle $\\\\theta$ with the positive $x$-axis such that $\\\\tan\\\\theta = 1$, what is the $y$-coordinate of point $P$?\",\n \"choices\": {\n \"A\": \"4\",\n \"B\": \"$4\\\\sqrt{2}$\",\n \"C\": \"8\",\n \"D\": \"$8\\\\sqrt{2}$\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"B\"\n },\n {\n \"id\": \"sat99-math-m1-q15\",\n \"number\": 15,\n \"type\": \"geometry_trigonometry\",\n \"prompt\": \"In right triangle $XYZ$, the measure of angle $Z$ is $90^{\\\\circ}$. If $\\\\sin(X) = \\\\frac{7}{25}$, what is the value of $\\\\cos(Y)$?\",\n \"choices\": {\n \"A\": \"$\\\\frac{24}{25}$\",\n \"B\": \"$\\\\frac{7}{24}$\",\n \"C\": \"$\\\\frac{25}{7}$\",\n \"D\": \"$\\\\frac{7}{25}$\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"D\"\n },\n {\n \"id\": \"sat99-math-m1-q16\",\n \"number\": 16,\n \"type\": \"geometry_trigonometry\",\n \"prompt\": \"A right circular cylinder is inscribed within a cube. The volume of the cube is 64 cubic centimeters. What is the volume of the cylinder, in cubic centimeters?\",\n \"choices\": {\n \"A\": \"$8\\\\pi$\",\n \"B\": \"$16\\\\pi$\",\n \"C\": \"$32\\\\pi$\",\n \"D\": \"$64\\\\pi$\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"B\"\n },\n {\n \"id\": \"sat99-math-m1-q17\",\n \"number\": 17,\n \"type\": \"geometry_trigonometry\",\n \"prompt\": \"A circle has a radius of 10. The area of a sector of the circle is $10\\\\pi$. What is the measure of the sector's central angle, in degrees?\",\n \"answer_format\": \"student_produced\",\n \"correct_answer\": \"36\"\n }\n]\n```" }, "status": "ok" }, { "attempts": 1, "index": 3, "output": { "costUsd": 0, "model": "gemini-2.5-pro", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 5 SAT Math questions (of which 1 are student-produced-response and the rest are MCQ).\nType: statistics_probability | Test: 99 | Unit: 1 | Start Q#: 18\n\nMANDATORY: Your JSON array MUST contain exactly 5 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nMCQ Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"statistics_probability\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\nSPR Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"statistics_probability\",\"prompt\":\"...\",\"answer_format\":\"student_produced\",\"correct_answer\":\"numeric_value\"}\nPlace SPR questions at the END of the array.\n\nFollow instructions for your type:\n\n[algebra] (MCQ + 1 SPR)\n- Topics: linear equations (1 var, 2 vars), systems of 2 linear equations, linear inequalities, absolute value\n- Mix word problems and abstract algebra. Use $...$ notation.\n- MCQ distractors: sign errors, arithmetic mistakes, setup errors, partial answers\n- SPR answers: clean integers or simple fractions\n\n[functions_analysis] (MCQ + 1 SPR)\n- Topics: function evaluation, domain/range, transformations (f(x)+k, f(x-h)), graphical analysis, composition f(g(x)), piecewise\n- At least one conceptual insight question (e.g., (p,q) and (q,p) on same graph)\n- Describe graphs in text. All computations yield clean results.\n\n[geometry_trigonometry] (MCQ + 1 SPR)\n- Topics: circle equations, Pythagorean theorem, special right triangles, coordinate geometry, trig ratios, volume/surface area, angle relationships\n- Optional \"visual\" field: {\"type\":\"coordinate_plane\",\"description\":\"...\"}\n- Radical answers OK (e.g., $5\\sqrt{13}$). Multi-step reasoning required.\n\n[statistics_probability] (MCQ + 1 SPR)\n- Topics: mean/median/mode, standard deviation (conceptual), probability (conditional from tables), data interpretation, scatter plots/line of best fit, sampling/bias\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Distractors: P(A|B) vs P(B|A) confusion, mean vs median, causal from correlational\n\n[advanced_math] (MCQ + 1 SPR)\n- Topics: quadratics (factoring, discriminant), polynomials (zeros, end behavior), exponentials (growth/decay), rational expressions, radical expressions, non-linear systems\n- Hardest questions on the test. Multi-step reasoning required. Elegant algebraic insight preferred.\n- All expressions must be factorable with integer coefficients. SPR answers are clean numbers.\n\nVERIFY EVERY ANSWER: Solve each problem step-by-step before finalizing correct_answer.\nOutput ONLY a valid JSON array of EXACTLY 5 objects. No markdown. Count them before outputting.", "text": "```json\n[\n {\n \"id\": \"sat99-math-m1-q18\",\n \"number\": 18,\n \"type\": \"statistics_probability\",\n \"prompt\": \"A scatterplot shows the relationship between the number of hours a student practices a musical instrument per week and the number of mistakes made during a performance. The line of best fit is given by the equation $y = -2.6x + 15$, where $y$ is the predicted number of mistakes and $x$ is the number of hours practiced. What is the best interpretation of the slope of the line of best fit?\",\n \"choices\": {\n \"A\": \"For each additional hour of practice, the predicted number of mistakes decreases by 2.6.\",\n \"B\": \"A student who does not practice is predicted to make 15 mistakes.\",\n \"C\": \"For every 2.6 additional mistakes made, the predicted number of practice hours decreases by 1.\",\n \"D\": \"The maximum possible number of mistakes is 15.\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"A\"\n },\n {\n \"id\": \"sat99-math-m1-q19\",\n \"number\": 19,\n \"type\": \"statistics_probability\",\n \"visual\": {\n \"type\": \"table\",\n \"title\": \"Participation in After-School Programs\",\n \"headers\": [\"\", \"Sports\", \"Arts\", \"Total\"],\n \"rows\": [\n [\"Juniors\", 85, 65, 150],\n [\"Seniors\", 95, 55, 150],\n [\"Total\", 180, 120, 300]\n ]\n },\n \"prompt\": \"The table above shows the distribution of 300 juniors and seniors at a high school based on their participation in after-school programs. If a student who participates in sports is selected at random, what is the probability that the student is a senior?\",\n \"choices\": {\n \"A\": \"$\\\\frac{95}{300}$\",\n \"B\": \"$\\\\frac{150}{300}$\",\n \"C\": \"$\\\\frac{95}{150}$\",\n \"D\": \"$\\\\frac{95}{180}$\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"D\"\n },\n {\n \"id\": \"sat99-math-m1-q20\",\n \"number\": 20,\n \"type\": \"statistics_probability\",\n \"prompt\": \"A data set consists of 9 integers. The numbers are 10, 12, 12, 13, 15, 17, 19, 20, and 45. Which of the following statements is true about the mean, median, and mode of this data set?\",\n \"choices\": {\n \"A\": \"The mean is greater than the median.\",\n \"B\": \"The median is greater than the mean.\",\n \"C\": \"The mode is greater than the mean.\",\n \"D\": \"The mean, median, and mode are all equal.\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"A\"\n },\n {\n \"id\": \"sat99-math-m1-q21\",\n \"number\": 21,\n \"type\": \"statistics_probability\",\n \"prompt\": \"An environmental scientist wants to estimate the proportion of trees in a large forest that are infected with a certain fungus. She divides the forest into 100 equally sized plots, randomly selects 10 plots, and examines every tree within those 10 plots. Which of the following best describes this sampling method?\",\n \"choices\": {\n \"A\": \"Simple random sampling\",\n \"B\": \"Cluster sampling\",\n \"C\": \"Stratified random sampling\",\n \"D\": \"Convenience sampling\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"B\"\n },\n {\n \"id\": \"sat99-math-m1-q22\",\n \"number\": 22,\n \"type\": \"statistics_probability\",\n \"visual\": {\n \"type\": \"table\",\n \"title\": \"Vehicle Types at a Service Center\",\n \"headers\": [\"\", \"Requires Oil Change\", \"Requires Tire Rotation\", \"Total\"],\n \"rows\": [\n [\"Sedan\", 50, 30, 80],\n [\"SUV\", 40, 35, 75],\n [\"Total\", 90, 65, 155]\n ]\n },\n \"prompt\": \"The table above shows the types of vehicles and the primary service requested at an auto service center on one day. Based on the table, what proportion of sedans required an oil change? (Express your answer as a decimal or fraction).\",\n \"answer_format\": \"student_produced\",\n \"correct_answer\": \"5/8\"\n }\n]\n```" }, "status": "ok" }, { "attempts": 2, "index": 4, "output": { "costUsd": 0, "model": "gemini-2.5-pro", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 5 SAT Math questions (of which 2 are student-produced-response and the rest are MCQ).\nType: advanced_math | Test: 99 | Unit: 1 | Start Q#: 23\n\nMANDATORY: Your JSON array MUST contain exactly 5 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nMCQ Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"advanced_math\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\nSPR Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"advanced_math\",\"prompt\":\"...\",\"answer_format\":\"student_produced\",\"correct_answer\":\"numeric_value\"}\nPlace SPR questions at the END of the array.\n\nFollow instructions for your type:\n\n[algebra] (MCQ + 2 SPR)\n- Topics: linear equations (1 var, 2 vars), systems of 2 linear equations, linear inequalities, absolute value\n- Mix word problems and abstract algebra. Use $...$ notation.\n- MCQ distractors: sign errors, arithmetic mistakes, setup errors, partial answers\n- SPR answers: clean integers or simple fractions\n\n[functions_analysis] (MCQ + 2 SPR)\n- Topics: function evaluation, domain/range, transformations (f(x)+k, f(x-h)), graphical analysis, composition f(g(x)), piecewise\n- At least one conceptual insight question (e.g., (p,q) and (q,p) on same graph)\n- Describe graphs in text. All computations yield clean results.\n\n[geometry_trigonometry] (MCQ + 2 SPR)\n- Topics: circle equations, Pythagorean theorem, special right triangles, coordinate geometry, trig ratios, volume/surface area, angle relationships\n- Optional \"visual\" field: {\"type\":\"coordinate_plane\",\"description\":\"...\"}\n- Radical answers OK (e.g., $5\\sqrt{13}$). Multi-step reasoning required.\n\n[statistics_probability] (MCQ + 2 SPR)\n- Topics: mean/median/mode, standard deviation (conceptual), probability (conditional from tables), data interpretation, scatter plots/line of best fit, sampling/bias\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Distractors: P(A|B) vs P(B|A) confusion, mean vs median, causal from correlational\n\n[advanced_math] (MCQ + 2 SPR)\n- Topics: quadratics (factoring, discriminant), polynomials (zeros, end behavior), exponentials (growth/decay), rational expressions, radical expressions, non-linear systems\n- Hardest questions on the test. Multi-step reasoning required. Elegant algebraic insight preferred.\n- All expressions must be factorable with integer coefficients. SPR answers are clean numbers.\n\nVERIFY EVERY ANSWER: Solve each problem step-by-step before finalizing correct_answer.\nOutput ONLY a valid JSON array of EXACTLY 5 objects. No markdown. Count them before outputting.", "text": "```json\n[\n {\n \"id\": \"sat99-math-m1-q23\",\n \"number\": 23,\n \"type\": \"advanced_math\",\n \"prompt\": \"The quadratic equation $kx^2 - 6x + 3 = 0$, where $k$ is a constant, has exactly one real solution. What is the value of $k$?\",\n \"choices\": {\n \"A\": \"-3\",\n \"B\": \"0\",\n \"C\": \"3\",\n \"D\": \"9\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"C\"\n },\n {\n \"id\": \"sat99-math-m1-q24\",\n \"number\": 24,\n \"type\": \"advanced_math\",\n \"prompt\": \"The polynomial $P(x) = x^3 - 2x^2 - 5x + 6$ has a zero at $x=1$. What is the sum of the other two zeros of the polynomial?\",\n \"choices\": {\n \"A\": \"-1\",\n \"B\": \"1\",\n \"C\": \"2\",\n \"D\": \"5\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"B\"\n },\n {\n \"id\": \"sat99-math-m1-q25\",\n \"number\": 25,\n \"type\": \"advanced_math\",\n \"prompt\": \"If $x > 2$, which of the following expressions is equivalent to $\\\\frac{x - \\\\frac{4}{x}}{1 + \\\\frac{2}{x}}$?\",\n \"choices\": {\n \"A\": \"x - 2\",\n \"B\": \"x + 2\",\n \"C\": \"\\\\frac{x-4}{3}\",\n \"D\": \"x - 4\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"A\"\n },\n {\n \"id\": \"sat99-math-m1-q26\",\n \"number\": 26,\n \"type\": \"advanced_math\",\n \"prompt\": \"A system of two equations is given by $y = x^2 - 5x + 7$ and $y - x = 2$. If $(x, y)$ is a solution to the system, what is one possible value for $x$?\",\n \"answer_format\": \"student_produced\",\n \"correct_answer\": \"5\"\n },\n {\n \"id\": \"sat99-math-m1-q27\",\n \"number\": 27,\n \"type\": \"advanced_math\",\n \"prompt\": \"If $n > 0$ and $\\\\sqrt[3]{n^2} \\\\cdot n^k = n^{5/3}$, what is the value of $k$?\",\n \"answer_format\": \"student_produced\",\n \"correct_answer\": \"1\"\n }\n]\n```" }, "status": "ok" }, { "attempts": 1, "index": 5, "output": { "costUsd": 0, "model": "gemini-2.5-pro", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 7 SAT Math questions (of which 2 are student-produced-response and the rest are MCQ).\nType: algebra | Test: 99 | Unit: 2 | Start Q#: 1\n\nMANDATORY: Your JSON array MUST contain exactly 7 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nMCQ Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"algebra\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\nSPR Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"algebra\",\"prompt\":\"...\",\"answer_format\":\"student_produced\",\"correct_answer\":\"numeric_value\"}\nPlace SPR questions at the END of the array.\n\nFollow instructions for your type:\n\n[algebra] (MCQ + 2 SPR)\n- Topics: linear equations (1 var, 2 vars), systems of 2 linear equations, linear inequalities, absolute value\n- Mix word problems and abstract algebra. Use $...$ notation.\n- MCQ distractors: sign errors, arithmetic mistakes, setup errors, partial answers\n- SPR answers: clean integers or simple fractions\n\n[functions_analysis] (MCQ + 2 SPR)\n- Topics: function evaluation, domain/range, transformations (f(x)+k, f(x-h)), graphical analysis, composition f(g(x)), piecewise\n- At least one conceptual insight question (e.g., (p,q) and (q,p) on same graph)\n- Describe graphs in text. All computations yield clean results.\n\n[geometry_trigonometry] (MCQ + 2 SPR)\n- Topics: circle equations, Pythagorean theorem, special right triangles, coordinate geometry, trig ratios, volume/surface area, angle relationships\n- Optional \"visual\" field: {\"type\":\"coordinate_plane\",\"description\":\"...\"}\n- Radical answers OK (e.g., $5\\sqrt{13}$). Multi-step reasoning required.\n\n[statistics_probability] (MCQ + 2 SPR)\n- Topics: mean/median/mode, standard deviation (conceptual), probability (conditional from tables), data interpretation, scatter plots/line of best fit, sampling/bias\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Distractors: P(A|B) vs P(B|A) confusion, mean vs median, causal from correlational\n\n[advanced_math] (MCQ + 2 SPR)\n- Topics: quadratics (factoring, discriminant), polynomials (zeros, end behavior), exponentials (growth/decay), rational expressions, radical expressions, non-linear systems\n- Hardest questions on the test. Multi-step reasoning required. Elegant algebraic insight preferred.\n- All expressions must be factorable with integer coefficients. SPR answers are clean numbers.\n\nVERIFY EVERY ANSWER: Solve each problem step-by-step before finalizing correct_answer.\nOutput ONLY a valid JSON array of EXACTLY 7 objects. No markdown. Count them before outputting.", "text": "```json\n[\n {\n \"id\": \"sat99-math-m2-q1\",\n \"number\": 1,\n \"type\": \"algebra\",\n \"prompt\": \"A car rental agency charges a flat daily fee of $40 plus an additional charge of $0.15 per mile driven. If a customer's total bill for a one-day rental was $98.50, how many miles did the customer drive?\",\n \"choices\": {\n \"A\": \"267\",\n \"B\": \"39\",\n \"C\": \"390\",\n \"D\": \"657\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"C\"\n },\n {\n \"id\": \"sat99-math-m2-q2\",\n \"number\": 2,\n \"type\": \"algebra\",\n \"prompt\": \"If $|2x - 3| = 11$, what is one possible value of $x$?\",\n \"choices\": {\n \"A\": \"7\",\n \"B\": \"4\",\n \"C\": \"-7\",\n \"D\": \"-4\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"D\"\n },\n {\n \"id\": \"sat99-math-m2-q3\",\n \"number\": 3,\n \"type\": \"algebra\",\n \"prompt\": \"At a food truck, 2 tacos and 1 burrito cost a total of $11.00. At the same food truck, 1 taco and 3 burritos cost a total of $18.00. What is the cost of one burrito?\",\n \"choices\": {\n \"A\": \"$2.50\",\n \"B\": \"$3.00\",\n \"C\": \"$5.50\",\n \"D\": \"$6.00\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"C\"\n },\n {\n \"id\": \"sat99-math-m2-q4\",\n \"number\": 4,\n \"type\": \"algebra\",\n \"prompt\": \"Which of the following values of $x$ is a solution to the inequality $8 - 3x < -7$?\",\n \"choices\": {\n \"A\": \"6\",\n \"B\": \"5\",\n \"C\": \"4\",\n \"D\": \"-5\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"A\"\n },\n {\n \"id\": \"sat99-math-m2-q5\",\n \"number\": 5,\n \"type\": \"algebra\",\n \"prompt\": \"A small business's weekly profit, $P$, in dollars, is modeled by the function $P(x) = 150x - 2700$, where $x$ represents the number of items sold. What is the minimum number of items the business must sell in a week to realize a profit?\",\n \"choices\": {\n \"A\": \"19\",\n \"B\": \"18\",\n \"C\": \"17\",\n \"D\": \"16\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"A\"\n },\n {\n \"id\": \"sat99-math-m2-q6\",\n \"number\": 6,\n \"type\": \"algebra\",\n \"prompt\": \"If $3x + 2y = 19$ and $x - y = 3$, what is the value of $x$?\",\n \"answer_format\": \"student_produced\",\n \"correct_answer\": \"5\"\n },\n {\n \"id\": \"sat99-math-m2-q7\",\n \"number\": 7,\n \"type\": \"algebra\",\n \"prompt\": \"If $\\\\frac{2}{3}(x - 4) = \\\\frac{1}{2}(x + 2)$, what is the value of $x$?\",\n \"answer_format\": \"student_produced\",\n \"correct_answer\": \"22\"\n }\n]\n```" }, "status": "ok" }, { "attempts": 1, "index": 6, "output": { "costUsd": 0, "model": "gemini-2.5-pro", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 5 SAT Math questions (of which 1 are student-produced-response and the rest are MCQ).\nType: functions_analysis | Test: 99 | Unit: 2 | Start Q#: 8\n\nMANDATORY: Your JSON array MUST contain exactly 5 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nMCQ Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"functions_analysis\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\nSPR Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"functions_analysis\",\"prompt\":\"...\",\"answer_format\":\"student_produced\",\"correct_answer\":\"numeric_value\"}\nPlace SPR questions at the END of the array.\n\nFollow instructions for your type:\n\n[algebra] (MCQ + 1 SPR)\n- Topics: linear equations (1 var, 2 vars), systems of 2 linear equations, linear inequalities, absolute value\n- Mix word problems and abstract algebra. Use $...$ notation.\n- MCQ distractors: sign errors, arithmetic mistakes, setup errors, partial answers\n- SPR answers: clean integers or simple fractions\n\n[functions_analysis] (MCQ + 1 SPR)\n- Topics: function evaluation, domain/range, transformations (f(x)+k, f(x-h)), graphical analysis, composition f(g(x)), piecewise\n- At least one conceptual insight question (e.g., (p,q) and (q,p) on same graph)\n- Describe graphs in text. All computations yield clean results.\n\n[geometry_trigonometry] (MCQ + 1 SPR)\n- Topics: circle equations, Pythagorean theorem, special right triangles, coordinate geometry, trig ratios, volume/surface area, angle relationships\n- Optional \"visual\" field: {\"type\":\"coordinate_plane\",\"description\":\"...\"}\n- Radical answers OK (e.g., $5\\sqrt{13}$). Multi-step reasoning required.\n\n[statistics_probability] (MCQ + 1 SPR)\n- Topics: mean/median/mode, standard deviation (conceptual), probability (conditional from tables), data interpretation, scatter plots/line of best fit, sampling/bias\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Distractors: P(A|B) vs P(B|A) confusion, mean vs median, causal from correlational\n\n[advanced_math] (MCQ + 1 SPR)\n- Topics: quadratics (factoring, discriminant), polynomials (zeros, end behavior), exponentials (growth/decay), rational expressions, radical expressions, non-linear systems\n- Hardest questions on the test. Multi-step reasoning required. Elegant algebraic insight preferred.\n- All expressions must be factorable with integer coefficients. SPR answers are clean numbers.\n\nVERIFY EVERY ANSWER: Solve each problem step-by-step before finalizing correct_answer.\nOutput ONLY a valid JSON array of EXACTLY 5 objects. No markdown. Count them before outputting.", "text": "```json\n[\n {\n \"id\": \"sat99-math-m2-q8\",\n \"number\": 8,\n \"type\": \"functions_analysis\",\n \"prompt\": \"For the function $f$ defined by $f(x) = 2x^2 - 5x + 1$, what is the value of $f(3) - f(-1)$?\",\n \"choices\": {\n \"A\": \"-4\",\n \"B\": \"-1\",\n \"C\": \"4\",\n \"D\": \"12\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"A\"\n },\n {\n \"id\": \"sat99-math-m2-q9\",\n \"number\": 9,\n \"type\": \"functions_analysis\",\n \"prompt\": \"The graph of a function $g$ is shown in the $xy$-plane. The function $h$ is defined by $h(x) = g(x-2) + 3$. If the point $(-1, 4)$ is on the graph of $g$, what is a point on the graph of $h$?\",\n \"choices\": {\n \"A\": \"(-3, 7)\",\n \"B\": \"(-3, 1)\",\n \"C\": \"(1, 7)\",\n \"D\": \"(1, 1)\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"C\"\n },\n {\n \"id\": \"sat99-math-m2-q10\",\n \"number\": 10,\n \"type\": \"functions_analysis\",\n \"prompt\": \"The function $f$ is an even function, meaning $f(x) = f(-x)$ for all values of $x$. The function $g$ is an odd function, meaning $g(-x) = -g(x)$ for all values of $x$. If $f(2) = 5$ and $g(-3) = -4$, what is the value of $f(-2) + g(3)$?\",\n \"choices\": {\n \"A\": \"-9\",\n \"B\": \"1\",\n \"C\": \"9\",\n \"D\": \"The value cannot be determined from the information given.\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"C\"\n },\n {\n \"id\": \"sat99-math-m2-q11\",\n \"number\": 11,\n \"type\": \"functions_analysis\",\n \"prompt\": \"Let the functions $p$ and $q$ be defined by $p(x) = x^2 - 4$ and $q(x) = 2x + 1$. For what positive value of $x$ is $p(q(x)) = 0$?\",\n \"choices\": {\n \"A\": \"$\\\\frac{1}{2}$\",\n \"B\": \"$\\\\frac{3}{2}$\",\n \"C\": \"$2$\",\n \"D\": \"$5$\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"A\"\n },\n {\n \"id\": \"sat99-math-m2-q12\",\n \"number\": 12,\n \"type\": \"functions_analysis\",\n \"prompt\": \"The function $f$ is defined below.\\n$f(x) = \\\\begin{cases} x^2+3 & \\\\text{if } x < 1 \\\\\\\\ 2x-1 & \\\\text{if } x \\\\ge 1 \\\\end{cases}$\\nWhat is the value of $f(f(0))$?\",\n \"answer_format\": \"student_produced\",\n \"correct_answer\": \"5\"\n }\n]\n```" }, "status": "ok" }, { "attempts": 1, "index": 7, "output": { "costUsd": 0, "model": "gemini-2.5-pro", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 5 SAT Math questions (of which 1 are student-produced-response and the rest are MCQ).\nType: geometry_trigonometry | Test: 99 | Unit: 2 | Start Q#: 13\n\nMANDATORY: Your JSON array MUST contain exactly 5 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nMCQ Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"geometry_trigonometry\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\nSPR Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"geometry_trigonometry\",\"prompt\":\"...\",\"answer_format\":\"student_produced\",\"correct_answer\":\"numeric_value\"}\nPlace SPR questions at the END of the array.\n\nFollow instructions for your type:\n\n[algebra] (MCQ + 1 SPR)\n- Topics: linear equations (1 var, 2 vars), systems of 2 linear equations, linear inequalities, absolute value\n- Mix word problems and abstract algebra. Use $...$ notation.\n- MCQ distractors: sign errors, arithmetic mistakes, setup errors, partial answers\n- SPR answers: clean integers or simple fractions\n\n[functions_analysis] (MCQ + 1 SPR)\n- Topics: function evaluation, domain/range, transformations (f(x)+k, f(x-h)), graphical analysis, composition f(g(x)), piecewise\n- At least one conceptual insight question (e.g., (p,q) and (q,p) on same graph)\n- Describe graphs in text. All computations yield clean results.\n\n[geometry_trigonometry] (MCQ + 1 SPR)\n- Topics: circle equations, Pythagorean theorem, special right triangles, coordinate geometry, trig ratios, volume/surface area, angle relationships\n- Optional \"visual\" field: {\"type\":\"coordinate_plane\",\"description\":\"...\"}\n- Radical answers OK (e.g., $5\\sqrt{13}$). Multi-step reasoning required.\n\n[statistics_probability] (MCQ + 1 SPR)\n- Topics: mean/median/mode, standard deviation (conceptual), probability (conditional from tables), data interpretation, scatter plots/line of best fit, sampling/bias\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Distractors: P(A|B) vs P(B|A) confusion, mean vs median, causal from correlational\n\n[advanced_math] (MCQ + 1 SPR)\n- Topics: quadratics (factoring, discriminant), polynomials (zeros, end behavior), exponentials (growth/decay), rational expressions, radical expressions, non-linear systems\n- Hardest questions on the test. Multi-step reasoning required. Elegant algebraic insight preferred.\n- All expressions must be factorable with integer coefficients. SPR answers are clean numbers.\n\nVERIFY EVERY ANSWER: Solve each problem step-by-step before finalizing correct_answer.\nOutput ONLY a valid JSON array of EXACTLY 5 objects. No markdown. Count them before outputting.", "text": "```json\n[\n {\n \"id\": \"sat99-math-m2-q13\",\n \"number\": 13,\n \"type\": \"geometry_trigonometry\",\n \"prompt\": \"A circle in the $xy$-plane has its center at $(3, -5)$. The point $(8, 7)$ lies on the circle. Which of the following is the equation of the circle?\",\n \"choices\": {\n \"A\": \"$(x-3)^2 + (y+5)^2 = 13$\",\n \"B\": \"$(x-8)^2 + (y-7)^2 = 169$\",\n \"C\": \"$(x+3)^2 + (y-5)^2 = 169$\",\n \"D\": \"$(x-3)^2 + (y+5)^2 = 169$\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"D\"\n },\n {\n \"id\": \"sat99-math-m2-q14\",\n \"number\": 14,\n \"type\": \"geometry_trigonometry\",\n \"prompt\": \"In right triangle $ABC$, the measure of angle $C$ is $90^{\\\\circ}$. If $\\\\sin(A) = \\\\frac{8}{17}$, what is the value of $\\\\cos(B)$?\",\n \"choices\": {\n \"A\": \"$\\\\frac{15}{17}$\",\n \"B\": \"$\\\\frac{8}{15}$\",\n \"C\": \"$\\\\frac{8}{17}$\",\n \"D\": \"$\\\\frac{17}{8}$\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"C\"\n },\n {\n \"id\": \"sat99-math-m2-q15\",\n \"number\": 15,\n \"type\": \"geometry_trigonometry\",\n \"prompt\": \"A cylindrical grain silo has a radius of 10 feet and a height of 40 feet. The top of the silo is a hemisphere with the same radius. What is the total volume of the silo in cubic feet?\",\n \"choices\": {\n \"A\": \"$4000\\\\pi$\",\n \"B\": \"$\\\\frac{14000}{3}\\\\pi$\",\n \"C\": \"$\\\\frac{16000}{3}\\\\pi$\",\n \"D\": \"$4400\\\\pi$\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"B\"\n },\n {\n \"id\": \"sat99-math-m2-q16\",\n \"number\": 16,\n \"type\": \"geometry_trigonometry\",\n \"prompt\": \"In the $xy$-plane, line $L$ is defined by the equation $y = -\\\\frac{1}{3}x + 5$. Line $M$ is perpendicular to line $L$ and passes through the point $(-2, 2)$. What is the $y$-intercept of line $M$?\",\n \"choices\": {\n \"A\": \"8\",\n \"B\": \"-8\",\n \"C\": \"3\",\n \"D\": \"$-\\\\frac{1}{3}$\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"A\"\n },\n {\n \"id\": \"sat99-math-m2-q17\",\n \"number\": 17,\n \"type\": \"geometry_trigonometry\",\n \"prompt\": \"A ladder is placed against a vertical wall. The base of the ladder is 7 feet away from the base of the wall, and the ladder makes an angle of $60^{\\\\circ}$ with the ground. What is the length of the ladder in feet?\",\n \"answer_format\": \"student_produced\",\n \"correct_answer\": \"14\"\n }\n]\n```" }, "status": "ok" }, { "attempts": 1, "index": 8, "output": { "costUsd": 0, "model": "gemini-2.5-pro", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 5 SAT Math questions (of which 1 are student-produced-response and the rest are MCQ).\nType: statistics_probability | Test: 99 | Unit: 2 | Start Q#: 18\n\nMANDATORY: Your JSON array MUST contain exactly 5 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nMCQ Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"statistics_probability\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\nSPR Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"statistics_probability\",\"prompt\":\"...\",\"answer_format\":\"student_produced\",\"correct_answer\":\"numeric_value\"}\nPlace SPR questions at the END of the array.\n\nFollow instructions for your type:\n\n[algebra] (MCQ + 1 SPR)\n- Topics: linear equations (1 var, 2 vars), systems of 2 linear equations, linear inequalities, absolute value\n- Mix word problems and abstract algebra. Use $...$ notation.\n- MCQ distractors: sign errors, arithmetic mistakes, setup errors, partial answers\n- SPR answers: clean integers or simple fractions\n\n[functions_analysis] (MCQ + 1 SPR)\n- Topics: function evaluation, domain/range, transformations (f(x)+k, f(x-h)), graphical analysis, composition f(g(x)), piecewise\n- At least one conceptual insight question (e.g., (p,q) and (q,p) on same graph)\n- Describe graphs in text. All computations yield clean results.\n\n[geometry_trigonometry] (MCQ + 1 SPR)\n- Topics: circle equations, Pythagorean theorem, special right triangles, coordinate geometry, trig ratios, volume/surface area, angle relationships\n- Optional \"visual\" field: {\"type\":\"coordinate_plane\",\"description\":\"...\"}\n- Radical answers OK (e.g., $5\\sqrt{13}$). Multi-step reasoning required.\n\n[statistics_probability] (MCQ + 1 SPR)\n- Topics: mean/median/mode, standard deviation (conceptual), probability (conditional from tables), data interpretation, scatter plots/line of best fit, sampling/bias\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Distractors: P(A|B) vs P(B|A) confusion, mean vs median, causal from correlational\n\n[advanced_math] (MCQ + 1 SPR)\n- Topics: quadratics (factoring, discriminant), polynomials (zeros, end behavior), exponentials (growth/decay), rational expressions, radical expressions, non-linear systems\n- Hardest questions on the test. Multi-step reasoning required. Elegant algebraic insight preferred.\n- All expressions must be factorable with integer coefficients. SPR answers are clean numbers.\n\nVERIFY EVERY ANSWER: Solve each problem step-by-step before finalizing correct_answer.\nOutput ONLY a valid JSON array of EXACTLY 5 objects. No markdown. Count them before outputting.", "text": "```json\n[\n {\n \"id\": \"sat99-math-m2-q18\",\n \"number\": 18,\n \"type\": \"statistics_probability\",\n \"prompt\": \"A market research firm surveyed 200 owners of either a sedan or an SUV. Each owner was asked whether their next vehicle purchase would be a new or used vehicle. The results are partially summarized in the table below. Of those surveyed, 60% owned a sedan.\",\n \"visual\": {\n \"type\": \"table\",\n \"title\": \"Next Vehicle Purchase Plan\",\n \"headers\": [\"Current Vehicle\", \"New\", \"Used\", \"Total\"],\n \"rows\": [\n [\"Sedan\", 72, \"\", \"\"],\n [\"SUV\", \"\", \"\", \"\"],\n [\"Total\", 110, 90, 200]\n ]\n },\n \"prompt_continue\": \"Based on the table, what is the probability that a person who currently owns an SUV will plan to purchase a new vehicle?\",\n \"choices\": {\n \"A\": \"0.34\",\n \"B\": \"0.475\",\n \"C\": \"0.525\",\n \"D\": \"0.65\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"B\"\n },\n {\n \"id\": \"sat99-math-m2-q19\",\n \"number\": 19,\n \"type\": \"statistics_probability\",\n \"prompt\": \"The dot plot above shows the number of goals scored by a soccer team in 15 games. If the game with the highest number of goals is removed from the data, which of the following statements about the mean and median of the resulting data is true?\",\n \"visual\": {\n \"type\": \"dot_plot\",\n \"description\": \"A dot plot titled 'Goals Scored in 15 Games'. The x-axis is labeled 'Number of Goals'. There are 3 dots at 0, 5 dots at 1, 4 dots at 2, 2 dots at 3, and 1 dot at 6.\"\n },\n \"choices\": {\n \"A\": \"The mean will decrease, and the median will remain the same.\",\n \"B\": \"The mean will decrease, and the median will decrease.\",\n \"C\": \"The mean will remain the same, and the median will decrease.\",\n \"D\": \"Both the mean and the median will remain the same.\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"A\"\n },\n {\n \"id\": \"sat99-math-m2-q20\",\n \"number\": 20,\n \"type\": \"statistics_probability\",\n \"prompt\": \"A sample of 50 trees was selected from a large forest. The height, in feet, and the trunk diameter, in inches, was measured for each tree. The data are shown in the scatterplot, along with a line of best fit. The equation of the line is $h = 3.8d + 15.2$, where $h$ is the height and $d$ is the trunk diameter.\",\n \"visual\": {\n \"type\": \"scatter_plot\",\n \"description\": \"A scatterplot titled 'Tree Height vs. Trunk Diameter' showing a positive linear association. The x-axis is 'Trunk Diameter (inches)' from 5 to 25. The y-axis is 'Height (feet)' from 20 to 120. A line of best fit with a positive slope is drawn through the data points.\"\n },\n \"prompt_continue\": \"Which of the following is the best interpretation of the slope of the line of best fit?\",\n \"choices\": {\n \"A\": \"For a tree with a trunk diameter of 0 inches, the predicted height is 3.8 feet.\",\n \"B\": \"For every 1-foot increase in a tree's height, the predicted trunk diameter increases by 3.8 inches.\",\n \"C\": \"For every 3.8-inch increase in a tree's trunk diameter, the predicted height increases by 1 foot.\",\n \"D\": \"For every 1-inch increase in a tree's trunk diameter, the predicted height increases by 3.8 feet.\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"D\"\n },\n {\n \"id\": \"sat99-math-m2-q21\",\n \"number\": 21,\n \"type\": \"statistics_probability\",\n \"prompt\": \"The test scores for two history classes are summarized by the box plots below.\",\n \"visual\": {\n \"type\": \"box_plot\",\n \"description\": \"Two box plots on a single axis from 50 to 100. 'Class A': min=55, Q1=65, median=75, Q3=80, max=90. 'Class B': min=60, Q1=75, median=80, Q3=85, max=90. The box for Class A is wider than the box for Class B, indicating a larger interquartile range.\"\n },\n \"prompt_continue\": \"Based on the box plots, which of the following must be true?\",\n \"choices\": {\n \"A\": \"The mean score of Class A is greater than the mean score of Class B.\",\n \"B\": \"The range of scores for Class A is greater than the range of scores for Class B.\",\n \"C\": \"The median score of Class B is greater than the median score of Class A.\",\n \"D\": \"There are more students in Class B than in Class A.\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"C\"\n },\n {\n \"id\": \"sat99-math-m2-q22\",\n \"number\": 22,\n \"type\": \"statistics_probability\",\n \"prompt\": \"A dataset consists of 7 distinct integers. The median of the dataset is 12, the minimum value is 5, and the range is 14. What is the maximum possible value of the mean of the dataset, rounded to the nearest tenth?\",\n \"answer_format\": \"student_produced\",\n \"correct_answer\": \"14.3\"\n }\n]\n```" }, "status": "ok" }, { "attempts": 1, "index": 9, "output": { "costUsd": 0, "model": "gemini-2.5-pro", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 5 SAT Math questions (of which 2 are student-produced-response and the rest are MCQ).\nType: advanced_math | Test: 99 | Unit: 2 | Start Q#: 23\n\nMANDATORY: Your JSON array MUST contain exactly 5 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nMCQ Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"advanced_math\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\nSPR Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"advanced_math\",\"prompt\":\"...\",\"answer_format\":\"student_produced\",\"correct_answer\":\"numeric_value\"}\nPlace SPR questions at the END of the array.\n\nFollow instructions for your type:\n\n[algebra] (MCQ + 2 SPR)\n- Topics: linear equations (1 var, 2 vars), systems of 2 linear equations, linear inequalities, absolute value\n- Mix word problems and abstract algebra. Use $...$ notation.\n- MCQ distractors: sign errors, arithmetic mistakes, setup errors, partial answers\n- SPR answers: clean integers or simple fractions\n\n[functions_analysis] (MCQ + 2 SPR)\n- Topics: function evaluation, domain/range, transformations (f(x)+k, f(x-h)), graphical analysis, composition f(g(x)), piecewise\n- At least one conceptual insight question (e.g., (p,q) and (q,p) on same graph)\n- Describe graphs in text. All computations yield clean results.\n\n[geometry_trigonometry] (MCQ + 2 SPR)\n- Topics: circle equations, Pythagorean theorem, special right triangles, coordinate geometry, trig ratios, volume/surface area, angle relationships\n- Optional \"visual\" field: {\"type\":\"coordinate_plane\",\"description\":\"...\"}\n- Radical answers OK (e.g., $5\\sqrt{13}$). Multi-step reasoning required.\n\n[statistics_probability] (MCQ + 2 SPR)\n- Topics: mean/median/mode, standard deviation (conceptual), probability (conditional from tables), data interpretation, scatter plots/line of best fit, sampling/bias\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Distractors: P(A|B) vs P(B|A) confusion, mean vs median, causal from correlational\n\n[advanced_math] (MCQ + 2 SPR)\n- Topics: quadratics (factoring, discriminant), polynomials (zeros, end behavior), exponentials (growth/decay), rational expressions, radical expressions, non-linear systems\n- Hardest questions on the test. Multi-step reasoning required. Elegant algebraic insight preferred.\n- All expressions must be factorable with integer coefficients. SPR answers are clean numbers.\n\nVERIFY EVERY ANSWER: Solve each problem step-by-step before finalizing correct_answer.\nOutput ONLY a valid JSON array of EXACTLY 5 objects. No markdown. Count them before outputting.", "text": "```json\n[\n {\n \"id\": \"sat99-math-m2-q23\",\n \"number\": 23,\n \"type\": \"advanced_math\",\n \"prompt\": \"The system of equations $y = x^2 - 4x + 7$ and $y = 2x + 2$ is graphed in the $xy$-plane. What is the sum of the $y$-coordinates of the two intersection points?\",\n \"choices\": {\n \"A\": \"6\",\n \"B\": \"12\",\n \"C\": \"18\",\n \"D\": \"24\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"C\"\n },\n {\n \"id\": \"sat99-math-m2-q24\",\n \"number\": 24,\n \"type\": \"advanced_math\",\n \"prompt\": \"A polynomial function $p(x)$ has zeros at $x=-1$ and $x=4$, where the zero at $x=4$ has a multiplicity of 2. The graph of $y=p(x)$ has a positive $y$-intercept. Which of the following could be the equation for $p(x)$?\",\n \"choices\": {\n \"A\": \"p(x) = (x+1)(x-4)^2\",\n \"B\": \"p(x) = (x-1)(x+4)^2\",\n \"C\": \"p(x) = -(x-1)(x+4)^2\",\n \"D\": \"p(x) = -(x+1)(x-4)^2\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"D\"\n },\n {\n \"id\": \"sat99-math-m2-q25\",\n \"number\": 25,\n \"type\": \"advanced_math\",\n \"prompt\": \"The function $f$ is defined by $f(x) = \\\\frac{2x^2 + 7x - 15}{x+5}$. The graph of $y=f(x)$ is a line with a point of discontinuity. What is the $y$-coordinate of the point of discontinuity?\",\n \"choices\": {\n \"A\": \"-13\",\n \"B\": \"-5\",\n \"C\": \"2\",\n \"D\": \"3\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"A\"\n },\n {\n \"id\": \"sat99-math-m2-q26\",\n \"number\": 26,\n \"type\": \"advanced_math\",\n \"prompt\": \"If $x > 0$, what value of $x$ is the solution to the equation $\\\\sqrt{2x+3} = x$?\",\n \"answer_format\": \"student_produced\",\n \"correct_answer\": \"3\"\n },\n {\n \"id\": \"sat99-math-m2-q27\",\n \"number\": 27,\n \"type\": \"advanced_math\",\n \"prompt\": \"The value of an investment is modeled by the function $V(t) = 100(1.5)^{\\\\frac{t}{4}}$, where $V(t)$ is the value of the investment in thousands of dollars and $t$ is the number of years since the initial investment was made. After how many years will the value of the investment be $225,000?\",\n \"answer_format\": \"student_produced\",\n \"correct_answer\": \"8\"\n }\n]\n```" }, "status": "ok" } ], "succeeded": 10 } }, "upstream": { "gen-rw": { "count": 14, "costUsd": 1.91406, "failed": 0, "items": [ { "costUsd": 0.08664, "model": "opus", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 5 SAT Reading & Writing questions.\nType: vocabulary_in_context | Test: 99 | Unit: 1 | Start Q#: 1\n\nMANDATORY: Your JSON array MUST contain exactly 5 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nFollow the instructions for your type:\n\n[vocabulary_in_context]\n- 2-3 sentence passage with ______ blank from specific academic domains (molecular biology, Renaissance art, behavioral economics, astrophysics, political philosophy)\n- 4 college-level word choices from the same semantic field\n- Correct word is most PRECISE for context, not most common synonym\n- ID format: sat{testNumber}-rw-m{mod}-q{N}\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"vocabulary_in_context\",\"passage\":\"...with ______...\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[text_comprehension]\n- 3-8 sentence passage from published-quality source (journal, novel, essay) with specific names/dates/findings\n- Mix subtypes: main idea (2-3), author's purpose (1-2), inference (2-3), tone (0-1), structure (1-2)\n- Topics: natural science, social science, literature, history, arts/humanities (use each)\n- Distractors: surface-level misreading, wrong scope, opposite interpretation\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"text_comprehension\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[data_analysis]\n- Passage describing realistic data/study results requiring INTERPRETATION, not lookup\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Data patterns: diminishing returns, optimal values, conditional relationships\n- Distractors: wrong column, wrong trend direction, correlation vs causation confusion\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"data_analysis\",\"passage\":\"...\",\"visual\":{...},\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[claim_support]\n- Mix: \"Which finding supports/undermines the claim?\" AND \"Which quotation illustrates the claim?\"\n- Claims must be specific and testable. Distractors are topically related but don't address the specific mechanism.\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"claim_support\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[standard_english_conventions]\n- Passage with [underlined: text here] portion. Student picks best replacement.\n- Cover: subject-verb agreement, pronouns, verb tense, parallel structure, modifiers, commas, semicolons, apostrophes, sentence boundaries\n- \"NO CHANGE\" (choice A) correct ~25% of the time, NOT always\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"standard_english_conventions\",\"passage\":\"...with [underlined: ...]...\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[transitions]\n- Passage with ______ blank connecting two ideas. All 4 choices grammatically correct; distinction is logical.\n- Cover: contrast, addition, cause/effect, example, sequence, concession\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"transitions\",\"passage\":\"...______ ...\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[rhetorical_synthesis]\n- Notes format: \"While researching [topic], a student has taken these notes:\\n- Note 1\\n- Note 2\\n- Note 3\"\n- Specific rhetorical goal stated. All 4 choices are grammatical sentences using the notes; only 1 achieves the goal.\n- Goals: emphasize difference, highlight finding, introduce topic, summarize conclusion, present cause-effect\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"rhetorical_synthesis\",\"passage\":\"...\",\"prompt\":\"The student wants to [goal]. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\nOutput ONLY a valid JSON array of EXACTLY 5 objects. No markdown. Count them before outputting.", "text": "[{\"id\":\"sat99-rw-m1-q1\",\"number\":1,\"type\":\"vocabulary_in_context\",\"passage\":\"In his 1503 treatise on proportion, the Florentine sculptor argued that classical statuary achieved its effects not through rigid adherence to mathematical ratios but through subtle departures from them. A slight elongation of the torso, a marginal tilt of the head—these deviations, he claimed, ______ the figure with a vitality that strict symmetry could never produce.\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"imbued\",\"B\":\"contaminated\",\"C\":\"stocked\",\"D\":\"appointed\"},\"answer_format\":\"mcq\",\"correct_answer\":\"A\"},{\"id\":\"sat99-rw-m1-q2\",\"number\":2,\"type\":\"vocabulary_in_context\",\"passage\":\"Behavioral economist Sendhil Mullainathan has argued that scarcity does not merely shape decisions—it ______ cognition itself. When the poor must constantly calculate trade-offs that the wealthy never notice, mental bandwidth is consumed by the arithmetic of survival, leaving fewer resources for long-term planning.\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"designates\",\"B\":\"restructures\",\"C\":\"transcribes\",\"D\":\"deputizes\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m1-q3\",\"number\":3,\"type\":\"vocabulary_in_context\",\"passage\":\"The discovery of FOXP2, often misleadingly dubbed the ''language gene,'' has been widely ______ in popular science writing. Though mutations in the gene do disrupt speech production in humans, researchers caution that it is one component of a vast regulatory network rather than a singular switch governing linguistic capacity.\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"repudiated\",\"B\":\"curtailed\",\"C\":\"oversimplified\",\"D\":\"fabricated\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m1-q4\",\"number\":4,\"type\":\"vocabulary_in_context\",\"passage\":\"Because Population III stars—the hypothetical first generation formed from primordial hydrogen and helium—have never been directly observed, their properties must be ______ from indirect evidence such as the chemical signatures of ancient, metal-poor stars still visible in the Milky Way's halo.\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"extracted\",\"B\":\"inferred\",\"C\":\"fabricated\",\"D\":\"disclosed\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m1-q5\",\"number\":5,\"type\":\"vocabulary_in_context\",\"passage\":\"In $\\\\textit{On Liberty}$, John Stuart Mill worried that majority opinion, when unchecked, could exert a form of coercion more insidious than any statute. Public disapproval, he argued, could ______ dissent so thoroughly that heterodox ideas would wither before they could even be articulated.\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"stifle\",\"B\":\"rebuke\",\"C\":\"dispute\",\"D\":\"evade\"},\"answer_format\":\"mcq\",\"correct_answer\":\"A\"}]" }, { "costUsd": 0.280965, "model": "opus", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 8 SAT Reading & Writing questions.\nType: text_comprehension | Test: 99 | Unit: 1 | Start Q#: 6\n\nMANDATORY: Your JSON array MUST contain exactly 8 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nFollow the instructions for your type:\n\n[vocabulary_in_context]\n- 2-3 sentence passage with ______ blank from specific academic domains (molecular biology, Renaissance art, behavioral economics, astrophysics, political philosophy)\n- 4 college-level word choices from the same semantic field\n- Correct word is most PRECISE for context, not most common synonym\n- ID format: sat{testNumber}-rw-m{mod}-q{N}\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"vocabulary_in_context\",\"passage\":\"...with ______...\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[text_comprehension]\n- 3-8 sentence passage from published-quality source (journal, novel, essay) with specific names/dates/findings\n- Mix subtypes: main idea (2-3), author's purpose (1-2), inference (2-3), tone (0-1), structure (1-2)\n- Topics: natural science, social science, literature, history, arts/humanities (use each)\n- Distractors: surface-level misreading, wrong scope, opposite interpretation\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"text_comprehension\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[data_analysis]\n- Passage describing realistic data/study results requiring INTERPRETATION, not lookup\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Data patterns: diminishing returns, optimal values, conditional relationships\n- Distractors: wrong column, wrong trend direction, correlation vs causation confusion\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"data_analysis\",\"passage\":\"...\",\"visual\":{...},\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[claim_support]\n- Mix: \"Which finding supports/undermines the claim?\" AND \"Which quotation illustrates the claim?\"\n- Claims must be specific and testable. Distractors are topically related but don't address the specific mechanism.\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"claim_support\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[standard_english_conventions]\n- Passage with [underlined: text here] portion. Student picks best replacement.\n- Cover: subject-verb agreement, pronouns, verb tense, parallel structure, modifiers, commas, semicolons, apostrophes, sentence boundaries\n- \"NO CHANGE\" (choice A) correct ~25% of the time, NOT always\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"standard_english_conventions\",\"passage\":\"...with [underlined: ...]...\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[transitions]\n- Passage with ______ blank connecting two ideas. All 4 choices grammatically correct; distinction is logical.\n- Cover: contrast, addition, cause/effect, example, sequence, concession\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"transitions\",\"passage\":\"...______ ...\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[rhetorical_synthesis]\n- Notes format: \"While researching [topic], a student has taken these notes:\\n- Note 1\\n- Note 2\\n- Note 3\"\n- Specific rhetorical goal stated. All 4 choices are grammatical sentences using the notes; only 1 achieves the goal.\n- Goals: emphasize difference, highlight finding, introduce topic, summarize conclusion, present cause-effect\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"rhetorical_synthesis\",\"passage\":\"...\",\"prompt\":\"The student wants to [goal]. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\nOutput ONLY a valid JSON array of EXACTLY 8 objects. No markdown. Count them before outputting.", "text": "[{\"id\":\"sat99-rw-m1-q6\",\"number\":6,\"type\":\"text_comprehension\",\"passage\":\"In her 2019 study of cooperative breeding in pied babblers (Turdoides bicolor), zoologist Amanda Ridley documented that subordinate adults who had previously received vocal encouragement from dominant pair members were significantly more likely to provision unrelated nestlings than those who had not. Ridley argues that this pattern complicates the standard kin-selection account, which holds that helping behaviors evolve primarily because they benefit close genetic relatives. The vocal exchanges, she suggests, function less as instructions than as a kind of social currency that helpers accumulate and later redeem through opportunities to feed and be associated with the brood.\",\"prompt\":\"Which choice best states the main idea of the text?\",\"choices\":{\"A\":\"Ridley's research demonstrates that kin selection plays no measurable role in the cooperative behavior of pied babblers.\",\"B\":\"Ridley's findings suggest that social interactions, rather than genetic relatedness alone, may shape helping behavior in pied babblers.\",\"C\":\"Ridley discovered that dominant pied babblers use vocal commands to direct unrelated subordinates toward specific nestlings.\",\"D\":\"Ridley concludes that subordinate pied babblers provision nestlings primarily to gain reproductive opportunities of their own.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m1-q7\",\"number\":7,\"type\":\"text_comprehension\",\"passage\":\"The following text is adapted from Edith Wharton's 1905 novel The House of Mirth. Lily Bart, a young woman of limited means, has just encountered an acquaintance at a train station.\\n\\nSelden paused in surprise. In the afternoon rush of the Grand Central Station his eyes had been refreshed by the sight of Miss Lily Bart. It was a Monday in early September, and he was returning to his work from a hurried dip into the country; but what was Miss Bart doing in town at that season? If she had appeared to be catching a train, he might have inferred that he had come on her in the act of transition between one and another of the country-houses which disputed her presence after the close of the Newport season; but her desultory air perplexed him. She stood apart from the crowd, letting it drift by her to the platform or the street, and wearing an air of irresolution which might, as he surmised, be the mask of a very definite purpose.\",\"prompt\":\"Which choice best describes the function of the final sentence in the text as a whole?\",\"choices\":{\"A\":\"It resolves Selden's earlier confusion by clarifying Lily's reasons for being at the station.\",\"B\":\"It contrasts Lily's outward behavior with the more chaotic behavior of the surrounding crowd.\",\"C\":\"It introduces Selden's suspicion that Lily's apparent indecision may conceal a deliberate intention.\",\"D\":\"It suggests that Lily is unaware of Selden's presence and is preoccupied with private worries.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m1-q8\",\"number\":8,\"type\":\"text_comprehension\",\"passage\":\"Art historian Patricia Fortini Brown has long maintained that fifteenth-century Venetian painters approached classical antiquity differently from their Florentine contemporaries. Where Florentine artists such as Mantegna treated ancient ruins and inscriptions as authoritative models to be reconstructed with archaeological precision, Venetian painters like Gentile Bellini incorporated antique fragments into compositions that subordinated historical accuracy to civic narrative. For Brown, the resulting images are not failures of scholarship but deliberate acts of appropriation: Venice, lacking a Roman past of its own, fashioned one by absorbing the relics of others into a visual rhetoric of imperial continuity.\",\"prompt\":\"As used in the text, the phrase \\\"deliberate acts of appropriation\\\" most nearly suggests that Venetian painters\",\"choices\":{\"A\":\"borrowed antique elements carelessly because they lacked access to reliable archaeological sources.\",\"B\":\"intentionally repurposed classical material to construct a flattering historical identity for their city.\",\"C\":\"competed with Florentine artists by producing more historically accurate reconstructions of Roman monuments.\",\"D\":\"rejected classical sources entirely in favor of subjects drawn from Venetian civic life.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m1-q9\",\"number\":9,\"type\":\"text_comprehension\",\"passage\":\"Recent measurements from the Parker Solar Probe have revealed that the solar wind is punctuated by sharp, S-shaped reversals in the magnetic field, which mission scientists have termed \\\"switchbacks.\\\" Although these structures had been glimpsed by earlier spacecraft, their abundance close to the Sun was not anticipated by prevailing models of coronal heating. Justin Kasper, the mission's principal investigator, cautions that it remains unclear whether switchbacks are generated near the solar surface by small-scale magnetic reconnection events or arise farther out through turbulent interactions in the expanding wind. Distinguishing between these possibilities, he notes, will require coordinated observations from probes operating at different distances from the Sun.\",\"prompt\":\"Based on the text, what can most reasonably be inferred about the origin of switchbacks?\",\"choices\":{\"A\":\"Researchers have determined that switchbacks form primarily through magnetic reconnection at the Sun's surface.\",\"B\":\"Researchers consider switchbacks too rare near the Sun to be explained by current coronal heating models.\",\"C\":\"Researchers do not yet have observations sufficient to decide between competing explanations of how switchbacks form.\",\"D\":\"Researchers believe switchbacks are an artifact of the Parker Solar Probe's instruments rather than a genuine solar phenomenon.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m1-q10\",\"number\":10,\"type\":\"text_comprehension\",\"passage\":\"In The Federalist No. 10, James Madison concedes that factions, which he defines as groups united by interests adverse to the rights of others or to the permanent interests of the community, are an unavoidable feature of free societies. To suppress factions, he argues, would require either destroying the liberty that nourishes them or imposing a uniformity of opinion that is impossible among reasoning beings. Madison therefore turns from the question of preventing factions to that of controlling their effects, contending that an extended republic, by multiplying the variety of interests it contains, makes the formation of an oppressive majority less probable than it would be in a small, homogeneous democracy.\",\"prompt\":\"Which choice best describes the overall structure of the text?\",\"choices\":{\"A\":\"It defines a political problem, dismisses two proposed solutions, and advances an alternative remedy.\",\"B\":\"It presents a historical example, identifies its weaknesses, and proposes a return to earlier practices.\",\"C\":\"It introduces a controversial claim, surveys objections to it, and concludes that the claim cannot be defended.\",\"D\":\"It compares two competing forms of government and recommends a compromise between them.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"A\"},{\"id\":\"sat99-rw-m1-q11\",\"number\":11,\"type\":\"text_comprehension\",\"passage\":\"The following text is adapted from a 1903 essay by W. E. B. Du Bois.\\n\\nBetween me and the other world there is ever an unasked question: unasked by some through feelings of delicacy; by others through the difficulty of rightly framing it. All, nevertheless, flutter round it. They approach me in a half-hesitant sort of way, eye me curiously or compassionately, and then, instead of saying directly, How does it feel to be a problem? they say, I know an excellent colored man in my town; or, I fought at Mechanicsville; or, Do not these Southern outrages make your blood boil? At these I smile, or am interested, or reduce the boiling to a simmer, as the occasion may require. To the real question, How does it feel to be a problem? I answer seldom a word.\",\"prompt\":\"Which choice best describes the speaker's tone in the text?\",\"choices\":{\"A\":\"Wry restraint mingled with quiet weariness\",\"B\":\"Defiant outrage barely contained by politeness\",\"C\":\"Cheerful detachment about a trivial misunderstanding\",\"D\":\"Hopeful conviction that misunderstandings will soon be resolved\"},\"answer_format\":\"mcq\",\"correct_answer\":\"A\"},{\"id\":\"sat99-rw-m1-q12\",\"number\":12,\"type\":\"text_comprehension\",\"passage\":\"Behavioral economist Sendhil Mullainathan and his collaborators have argued that the experience of scarcity, whether of money, time, or calories, alters cognition in ways that conventional economic models fail to capture. In a series of experiments, participants asked to manage tight budgets performed better on tasks directly related to those budgets but worse on unrelated tests of fluid intelligence and self-control. Mullainathan describes this trade-off as a kind of cognitive \\\"tax\\\" levied by scarcity itself, one that may help explain why interventions designed to alleviate poverty often falter when they demand additional planning, paperwork, or attention from the very people they aim to assist.\",\"prompt\":\"Why does the author most likely describe scarcity as a cognitive \\\"tax\\\"?\",\"choices\":{\"A\":\"To suggest that governments should compensate individuals financially for the mental costs of being poor\",\"B\":\"To emphasize that scarcity imposes mental costs that can undermine programs requiring sustained effort from beneficiaries\",\"C\":\"To argue that conventional economic models accurately predict how people respond to financial constraints\",\"D\":\"To imply that people experiencing scarcity perform worse on every cognitive task than wealthier people do\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m1-q13\",\"number\":13,\"type\":\"text_comprehension\",\"passage\":\"Marine biologist Roger Payne's 1971 recordings of humpback whale songs are often credited with galvanizing public support for the international moratorium on commercial whaling adopted a decade later. Yet the recordings' influence cannot be attributed to scientific argument alone. By releasing the songs as a commercial album rather than confining them to academic journals, Payne placed the voices of whales within the same cultural space as popular music, inviting listeners to perceive the animals as expressive beings rather than as units of harvestable biomass. The conservation movement that followed drew as much from this aesthetic reframing as from the demographic data that researchers were simultaneously compiling.\",\"prompt\":\"Which choice best states the main purpose of the text?\",\"choices\":{\"A\":\"To document the scientific methods Payne used to record and analyze humpback whale vocalizations\",\"B\":\"To argue that the moratorium on commercial whaling was adopted primarily because of advances in population data\",\"C\":\"To suggest that Payne's recordings shaped public attitudes toward whales in part by reframing them aesthetically rather than only scientifically\",\"D\":\"To criticize the conservation movement for relying on emotional appeals rather than on rigorous scientific evidence\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"}]" }, { "costUsd": 0.15984, "model": "opus", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 4 SAT Reading & Writing questions.\nType: data_analysis | Test: 99 | Unit: 1 | Start Q#: 14\n\nMANDATORY: Your JSON array MUST contain exactly 4 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nFollow the instructions for your type:\n\n[vocabulary_in_context]\n- 2-3 sentence passage with ______ blank from specific academic domains (molecular biology, Renaissance art, behavioral economics, astrophysics, political philosophy)\n- 4 college-level word choices from the same semantic field\n- Correct word is most PRECISE for context, not most common synonym\n- ID format: sat{testNumber}-rw-m{mod}-q{N}\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"vocabulary_in_context\",\"passage\":\"...with ______...\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[text_comprehension]\n- 3-8 sentence passage from published-quality source (journal, novel, essay) with specific names/dates/findings\n- Mix subtypes: main idea (2-3), author's purpose (1-2), inference (2-3), tone (0-1), structure (1-2)\n- Topics: natural science, social science, literature, history, arts/humanities (use each)\n- Distractors: surface-level misreading, wrong scope, opposite interpretation\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"text_comprehension\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[data_analysis]\n- Passage describing realistic data/study results requiring INTERPRETATION, not lookup\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Data patterns: diminishing returns, optimal values, conditional relationships\n- Distractors: wrong column, wrong trend direction, correlation vs causation confusion\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"data_analysis\",\"passage\":\"...\",\"visual\":{...},\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[claim_support]\n- Mix: \"Which finding supports/undermines the claim?\" AND \"Which quotation illustrates the claim?\"\n- Claims must be specific and testable. Distractors are topically related but don't address the specific mechanism.\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"claim_support\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[standard_english_conventions]\n- Passage with [underlined: text here] portion. Student picks best replacement.\n- Cover: subject-verb agreement, pronouns, verb tense, parallel structure, modifiers, commas, semicolons, apostrophes, sentence boundaries\n- \"NO CHANGE\" (choice A) correct ~25% of the time, NOT always\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"standard_english_conventions\",\"passage\":\"...with [underlined: ...]...\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[transitions]\n- Passage with ______ blank connecting two ideas. All 4 choices grammatically correct; distinction is logical.\n- Cover: contrast, addition, cause/effect, example, sequence, concession\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"transitions\",\"passage\":\"...______ ...\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[rhetorical_synthesis]\n- Notes format: \"While researching [topic], a student has taken these notes:\\n- Note 1\\n- Note 2\\n- Note 3\"\n- Specific rhetorical goal stated. All 4 choices are grammatical sentences using the notes; only 1 achieves the goal.\n- Goals: emphasize difference, highlight finding, introduce topic, summarize conclusion, present cause-effect\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"rhetorical_synthesis\",\"passage\":\"...\",\"prompt\":\"The student wants to [goal]. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\nOutput ONLY a valid JSON array of EXACTLY 4 objects. No markdown. Count them before outputting.", "text": "[{\"id\":\"sat99-rw-m1-q14\",\"number\":14,\"type\":\"data_analysis\",\"passage\":\"Soil ecologist Maren Holst investigated how earthworm density affects the decomposition rate of leaf litter in temperate deciduous forests. She established experimental plots with varying numbers of earthworms per square meter and measured the percentage of leaf mass lost over a 90-day period. Holst hypothesized that decomposition would increase with earthworm density but that the rate of increase would diminish at higher densities because of resource competition among the worms.\",\"visual\":{\"type\":\"table\",\"title\":\"Leaf Litter Decomposition by Earthworm Density (90 Days)\",\"headers\":[\"Earthworms per m²\",\"Leaf Mass Lost (%)\",\"Increase from Previous Level (%)\"],\"rows\":[[\"0\",\"12\",\"—\"],[\"20\",\"28\",\"16\"],[\"40\",\"41\",\"13\"],[\"60\",\"49\",\"8\"],[\"80\",\"53\",\"4\"]]},\"prompt\":\"Which choice best describes data from the table that support Holst's hypothesis?\",\"choices\":{\"A\":\"Decomposition was highest at 80 earthworms per m², and the increase from the previous density level was the smallest observed.\",\"B\":\"Decomposition increased steadily from 0 to 80 earthworms per m², and the increase from each previous level was identical across densities.\",\"C\":\"Decomposition was lowest at 0 earthworms per m², and the increase from 0 to 20 earthworms per m² was smaller than the increase from 60 to 80.\",\"D\":\"Decomposition peaked at 40 earthworms per m², after which leaf mass lost declined as density increased further.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"A\"},{\"id\":\"sat99-rw-m1-q15\",\"number\":15,\"type\":\"data_analysis\",\"passage\":\"A team of behavioral economists led by Rajiv Chandra examined whether the timing of a charitable appeal influences donation amounts. Participants were randomly assigned to receive a solicitation email either on the morning of payday, three days after payday, or ten days after payday. The researchers recorded the mean donation in each condition as well as the percentage of recipients who donated at all.\",\"visual\":{\"type\":\"table\",\"title\":\"Donation Behavior by Email Timing\",\"headers\":[\"Timing of Appeal\",\"Donation Rate (%)\",\"Mean Donation ($)\"],\"rows\":[[\"Morning of payday\",\"34\",\"22.50\"],[\"3 days after payday\",\"29\",\"18.75\"],[\"10 days after payday\",\"18\",\"11.20\"]]},\"prompt\":\"Which choice most accurately describes the relationship between appeal timing and donor behavior shown in the table?\",\"choices\":{\"A\":\"As the delay after payday increased, the donation rate rose while the mean donation amount fell.\",\"B\":\"As the delay after payday increased, both the donation rate and the mean donation amount declined.\",\"C\":\"The donation rate was highest ten days after payday, but the mean donation was highest on the morning of payday.\",\"D\":\"Appeals sent three days after payday produced both the highest donation rate and the highest mean donation.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m1-q16\",\"number\":16,\"type\":\"data_analysis\",\"passage\":\"Materials scientist Inés Beltrán tested the tensile strength of a novel bio-based polymer at several curing temperatures to determine an optimal processing condition. She noted that insufficient curing leaves polymer chains weakly bonded, whereas excessive heat can degrade the chains and reduce strength. Beltrán predicted that tensile strength would rise with temperature up to a threshold and then fall beyond it.\",\"visual\":{\"type\":\"table\",\"title\":\"Tensile Strength of Bio-Based Polymer by Curing Temperature\",\"headers\":[\"Curing Temperature (°C)\",\"Tensile Strength (MPa)\"],\"rows\":[[\"100\",\"42\"],[\"120\",\"58\"],[\"140\",\"71\"],[\"160\",\"65\"],[\"180\",\"49\"]]},\"prompt\":\"Which choice best describes data from the table that support Beltrán's prediction?\",\"choices\":{\"A\":\"Tensile strength rose continuously from 100 °C to 180 °C, confirming that higher curing temperatures always improve the polymer.\",\"B\":\"Tensile strength was lowest at 100 °C and highest at 180 °C, indicating that thermal degradation did not occur within the tested range.\",\"C\":\"Tensile strength increased from 100 °C to 140 °C and then decreased at 160 °C and 180 °C, indicating a peak within the tested range.\",\"D\":\"Tensile strength remained constant between 120 °C and 160 °C, suggesting that curing temperature has little effect in that interval.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m1-q17\",\"number\":17,\"type\":\"data_analysis\",\"passage\":\"In a study of cognitive load during multitasking, psychologist Theo Abedi asked participants to complete a primary reading comprehension task while performing a secondary task of varying difficulty. He measured reading accuracy (percentage of comprehension questions answered correctly) and reaction time on the secondary task. Abedi wanted to determine whether a moderately demanding secondary task would disrupt primary-task performance less than either a very easy or a very difficult one.\",\"visual\":{\"type\":\"table\",\"title\":\"Performance by Secondary-Task Difficulty\",\"headers\":[\"Secondary-Task Difficulty\",\"Reading Accuracy (%)\",\"Secondary-Task Reaction Time (ms)\"],\"rows\":[[\"Very easy\",\"74\",\"310\"],[\"Moderate\",\"88\",\"425\"],[\"Difficult\",\"61\",\"690\"]]},\"prompt\":\"Which choice best describes data from the table that are relevant to Abedi's research question?\",\"choices\":{\"A\":\"Reading accuracy was lower in the difficult condition than in the very easy condition, showing that harder secondary tasks always impair primary-task performance.\",\"B\":\"Reading accuracy was higher in the moderate condition than in either the very easy or the difficult condition, consistent with the idea that a moderately demanding secondary task disrupts reading least.\",\"C\":\"Reaction time on the secondary task was shortest in the very easy condition, indicating that participants focused most effectively on reading when the secondary task was simplest.\",\"D\":\"Reading accuracy and secondary-task reaction time both increased as difficulty rose, suggesting that the two measures track each other across conditions.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"}]" }, { "costUsd": 0.150315, "model": "opus", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 4 SAT Reading & Writing questions.\nType: claim_support | Test: 99 | Unit: 1 | Start Q#: 18\n\nMANDATORY: Your JSON array MUST contain exactly 4 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nFollow the instructions for your type:\n\n[vocabulary_in_context]\n- 2-3 sentence passage with ______ blank from specific academic domains (molecular biology, Renaissance art, behavioral economics, astrophysics, political philosophy)\n- 4 college-level word choices from the same semantic field\n- Correct word is most PRECISE for context, not most common synonym\n- ID format: sat{testNumber}-rw-m{mod}-q{N}\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"vocabulary_in_context\",\"passage\":\"...with ______...\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[text_comprehension]\n- 3-8 sentence passage from published-quality source (journal, novel, essay) with specific names/dates/findings\n- Mix subtypes: main idea (2-3), author's purpose (1-2), inference (2-3), tone (0-1), structure (1-2)\n- Topics: natural science, social science, literature, history, arts/humanities (use each)\n- Distractors: surface-level misreading, wrong scope, opposite interpretation\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"text_comprehension\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[data_analysis]\n- Passage describing realistic data/study results requiring INTERPRETATION, not lookup\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Data patterns: diminishing returns, optimal values, conditional relationships\n- Distractors: wrong column, wrong trend direction, correlation vs causation confusion\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"data_analysis\",\"passage\":\"...\",\"visual\":{...},\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[claim_support]\n- Mix: \"Which finding supports/undermines the claim?\" AND \"Which quotation illustrates the claim?\"\n- Claims must be specific and testable. Distractors are topically related but don't address the specific mechanism.\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"claim_support\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[standard_english_conventions]\n- Passage with [underlined: text here] portion. Student picks best replacement.\n- Cover: subject-verb agreement, pronouns, verb tense, parallel structure, modifiers, commas, semicolons, apostrophes, sentence boundaries\n- \"NO CHANGE\" (choice A) correct ~25% of the time, NOT always\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"standard_english_conventions\",\"passage\":\"...with [underlined: ...]...\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[transitions]\n- Passage with ______ blank connecting two ideas. All 4 choices grammatically correct; distinction is logical.\n- Cover: contrast, addition, cause/effect, example, sequence, concession\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"transitions\",\"passage\":\"...______ ...\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[rhetorical_synthesis]\n- Notes format: \"While researching [topic], a student has taken these notes:\\n- Note 1\\n- Note 2\\n- Note 3\"\n- Specific rhetorical goal stated. All 4 choices are grammatical sentences using the notes; only 1 achieves the goal.\n- Goals: emphasize difference, highlight finding, introduce topic, summarize conclusion, present cause-effect\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"rhetorical_synthesis\",\"passage\":\"...\",\"prompt\":\"The student wants to [goal]. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\nOutput ONLY a valid JSON array of EXACTLY 4 objects. No markdown. Count them before outputting.", "text": "[{\"id\":\"sat99-rw-m1-q18\",\"number\":18,\"type\":\"claim_support\",\"passage\":\"A researcher studying the cognitive development of corvids (a family of birds including crows and ravens) has proposed that New Caledonian crows possess a form of causal reasoning that extends beyond simple trial-and-error learning. Specifically, the researcher claims that these birds can mentally simulate the physical consequences of an action before executing it, allowing them to solve novel tool-use problems on the first attempt without prior exposure.\",\"prompt\":\"Which finding, if true, would most directly support the researcher's claim?\",\"choices\":{\"A\":\"In laboratory settings, New Caledonian crows that observed other crows retrieving food with hooked sticks were subsequently able to use similar tools themselves after several practice sessions.\",\"B\":\"When presented with a novel apparatus requiring them to drop stones into a water-filled tube to raise a floating reward, naive New Caledonian crows selected stones of appropriate size and density on their first attempt, ignoring objects that would float.\",\"C\":\"New Caledonian crows in the wild have been documented manufacturing hooked tools from pandanus leaves, a behavior that varies regionally and appears to be transmitted across generations.\",\"D\":\"Neuroanatomical studies reveal that New Caledonian crows possess a higher density of neurons in the pallium than other corvid species of comparable body size.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m1-q19\",\"number\":19,\"type\":\"claim_support\",\"passage\":\"In her 2021 study of urban microclimates, geographer Lena Abernathy argues that the cooling effect of small neighborhood parks on surrounding city blocks is disproportionately large relative to their area—a phenomenon she attributes not to the parks' vegetation alone but to the way they disrupt the continuous surface of heat-absorbing pavement and buildings that characterize dense urban cores.\",\"prompt\":\"Which finding from Abernathy's study would most directly support her specific explanation for the cooling effect?\",\"choices\":{\"A\":\"Neighborhood parks measuring less than one hectare were associated with average temperature reductions of 1.8°C in adjacent blocks during summer afternoons.\",\"B\":\"Tree-canopy coverage within the parks correlated positively with the magnitude of temperature reduction observed in surrounding areas.\",\"C\":\"Paved plazas of equivalent size to the parks, when inserted into dense urban blocks, produced measurable cooling effects roughly 60 percent as large as those of vegetated parks.\",\"D\":\"Residents living within 200 meters of a neighborhood park reported lower rates of heat-related illness than residents living farther away during extreme heat events.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m1-q20\",\"number\":20,\"type\":\"claim_support\",\"passage\":\"In his critical study of the Victorian novel, literary scholar Marcus Holbrook contends that Elizabeth Gaskell's \\\"North and South\\\" (1855) departs from the conventions of the industrial novel by granting its female protagonist, Margaret Hale, a role as economic arbiter rather than as a passive moral influence. Holbrook argues that Margaret's decisions shape the material fortunes of the novel's central mill, a reversal of the typical Victorian pattern in which women inspire male action without directly determining business outcomes.\",\"prompt\":\"Which quotation from \\\"North and South\\\" would most effectively illustrate Holbrook's claim?\",\"choices\":{\"A\":\"\\\"Margaret felt, as she watched the crowds streaming from the mill gates, a sorrow she could scarcely name, a pity for the weary faces that seemed to reproach her own ease.\\\"\",\"B\":\"\\\"Mr. Thornton's manner toward her had grown gentler since their last conversation, and she was aware that her words had, in some measure, softened his views on the conduct of his workmen.\\\"\",\"C\":\"\\\"It was Margaret's capital, inherited from Mr. Bell and offered upon terms of her own devising, that permitted Thornton to reopen Marlborough Mills when no bank in Milton would advance him a shilling.\\\"\",\"D\":\"\\\"She thought often of the North, of its smoke and clangor, and wondered whether the gentleness of the South had not unfitted her for a world in which men like Thornton labored and contended.\\\"\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m1-q21\",\"number\":21,\"type\":\"claim_support\",\"passage\":\"Archaeologist Priya Chandran has argued that the residents of Çatalhöyük, a Neolithic settlement in southern Anatolia occupied from roughly 7100 to 5950 BCE, maintained a remarkably egalitarian social structure throughout the settlement's occupation, with no single household accumulating significantly greater wealth or status than its neighbors. Critics have suggested that this apparent egalitarianism may reflect the limitations of the archaeological record rather than the actual social organization of the community.\",\"prompt\":\"Which finding, if true, would most directly undermine Chandran's claim?\",\"choices\":{\"A\":\"Isotopic analysis of human remains from Çatalhöyük indicates that dietary composition varied little across individuals, suggesting broadly similar access to food resources.\",\"B\":\"Excavations at a contemporaneous settlement 200 kilometers to the east have revealed clear evidence of status differentiation, including elite burials with imported obsidian goods.\",\"C\":\"Detailed analysis of the distribution of rare imported materials across Çatalhöyük households reveals that a small cluster of dwellings consistently contained ten times the quantity found elsewhere in the settlement.\",\"D\":\"The layout of Çatalhöyük, with its densely packed houses sharing walls and accessed through roof openings, has been interpreted as evidence of strong communal identity.\",\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"}]" }, { "costUsd": 0.08124, "model": "opus", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 5 SAT Reading & Writing questions.\nType: standard_english_conventions | Test: 99 | Unit: 1 | Start Q#: 22\n\nMANDATORY: Your JSON array MUST contain exactly 5 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nFollow the instructions for your type:\n\n[vocabulary_in_context]\n- 2-3 sentence passage with ______ blank from specific academic domains (molecular biology, Renaissance art, behavioral economics, astrophysics, political philosophy)\n- 4 college-level word choices from the same semantic field\n- Correct word is most PRECISE for context, not most common synonym\n- ID format: sat{testNumber}-rw-m{mod}-q{N}\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"vocabulary_in_context\",\"passage\":\"...with ______...\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[text_comprehension]\n- 3-8 sentence passage from published-quality source (journal, novel, essay) with specific names/dates/findings\n- Mix subtypes: main idea (2-3), author's purpose (1-2), inference (2-3), tone (0-1), structure (1-2)\n- Topics: natural science, social science, literature, history, arts/humanities (use each)\n- Distractors: surface-level misreading, wrong scope, opposite interpretation\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"text_comprehension\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[data_analysis]\n- Passage describing realistic data/study results requiring INTERPRETATION, not lookup\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Data patterns: diminishing returns, optimal values, conditional relationships\n- Distractors: wrong column, wrong trend direction, correlation vs causation confusion\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"data_analysis\",\"passage\":\"...\",\"visual\":{...},\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[claim_support]\n- Mix: \"Which finding supports/undermines the claim?\" AND \"Which quotation illustrates the claim?\"\n- Claims must be specific and testable. Distractors are topically related but don't address the specific mechanism.\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"claim_support\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[standard_english_conventions]\n- Passage with [underlined: text here] portion. Student picks best replacement.\n- Cover: subject-verb agreement, pronouns, verb tense, parallel structure, modifiers, commas, semicolons, apostrophes, sentence boundaries\n- \"NO CHANGE\" (choice A) correct ~25% of the time, NOT always\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"standard_english_conventions\",\"passage\":\"...with [underlined: ...]...\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[transitions]\n- Passage with ______ blank connecting two ideas. All 4 choices grammatically correct; distinction is logical.\n- Cover: contrast, addition, cause/effect, example, sequence, concession\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"transitions\",\"passage\":\"...______ ...\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[rhetorical_synthesis]\n- Notes format: \"While researching [topic], a student has taken these notes:\\n- Note 1\\n- Note 2\\n- Note 3\"\n- Specific rhetorical goal stated. All 4 choices are grammatical sentences using the notes; only 1 achieves the goal.\n- Goals: emphasize difference, highlight finding, introduce topic, summarize conclusion, present cause-effect\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"rhetorical_synthesis\",\"passage\":\"...\",\"prompt\":\"The student wants to [goal]. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\nOutput ONLY a valid JSON array of EXACTLY 5 objects. No markdown. Count them before outputting.", "text": "[{\"id\":\"sat99-rw-m1-q22\",\"number\":22,\"type\":\"standard_english_conventions\",\"passage\":\"The sculptor Louise Nevelson, who emigrated from what is now Ukraine to the United States in 1905, became renowned for her monochromatic wooden assemblages; these works, often painted entirely black or white, [underlined: transforms] discarded objects into monumental compositions.\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"transform\",\"C\":\"is transforming\",\"D\":\"transforming\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m1-q23\",\"number\":23,\"type\":\"standard_english_conventions\",\"passage\":\"Marine biologist Sylvia Earle has spent decades advocating for ocean conservation. Her research expeditions have taken her to depths few humans have reached, and [underlined: they're findings have] reshaped public understanding of deep-sea ecosystems.\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"their findings have\",\"C\":\"there findings have\",\"D\":\"they're findings has\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m1-q24\",\"number\":24,\"type\":\"standard_english_conventions\",\"passage\":\"In traditional Japanese $kintsugi$, artisans repair broken pottery with lacquer mixed with powdered gold, silver, or [underlined: platinum, the technique treats] breakage and repair as part of an object's history rather than something to disguise.\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"platinum the technique treats\",\"C\":\"platinum; the technique treats\",\"D\":\"platinum, treating\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m1-q25\",\"number\":25,\"type\":\"standard_english_conventions\",\"passage\":\"Economist Esther Duflo, who shared the 2019 Nobel Prize for her experimental approach to alleviating global poverty, argues that small-scale randomized trials can reveal which interventions actually [underlined: works] in specific local contexts.\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"is working\",\"C\":\"has worked\",\"D\":\"work\"},\"answer_format\":\"mcq\",\"correct_answer\":\"D\"},{\"id\":\"sat99-rw-m1-q26\",\"number\":26,\"type\":\"standard_english_conventions\",\"passage\":\"By the time astronomers confirmed the existence of the first exoplanet orbiting a sun-like star in 1995, decades of speculation about worlds beyond our solar [underlined: system, had] given way to a concrete, observable field of research.\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"system had\",\"C\":\"system; had\",\"D\":\"system: had\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"}]" }, { "costUsd": 0.07734, "model": "opus", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 4 SAT Reading & Writing questions.\nType: transitions | Test: 99 | Unit: 1 | Start Q#: 27\n\nMANDATORY: Your JSON array MUST contain exactly 4 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nFollow the instructions for your type:\n\n[vocabulary_in_context]\n- 2-3 sentence passage with ______ blank from specific academic domains (molecular biology, Renaissance art, behavioral economics, astrophysics, political philosophy)\n- 4 college-level word choices from the same semantic field\n- Correct word is most PRECISE for context, not most common synonym\n- ID format: sat{testNumber}-rw-m{mod}-q{N}\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"vocabulary_in_context\",\"passage\":\"...with ______...\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[text_comprehension]\n- 3-8 sentence passage from published-quality source (journal, novel, essay) with specific names/dates/findings\n- Mix subtypes: main idea (2-3), author's purpose (1-2), inference (2-3), tone (0-1), structure (1-2)\n- Topics: natural science, social science, literature, history, arts/humanities (use each)\n- Distractors: surface-level misreading, wrong scope, opposite interpretation\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"text_comprehension\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[data_analysis]\n- Passage describing realistic data/study results requiring INTERPRETATION, not lookup\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Data patterns: diminishing returns, optimal values, conditional relationships\n- Distractors: wrong column, wrong trend direction, correlation vs causation confusion\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"data_analysis\",\"passage\":\"...\",\"visual\":{...},\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[claim_support]\n- Mix: \"Which finding supports/undermines the claim?\" AND \"Which quotation illustrates the claim?\"\n- Claims must be specific and testable. Distractors are topically related but don't address the specific mechanism.\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"claim_support\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[standard_english_conventions]\n- Passage with [underlined: text here] portion. Student picks best replacement.\n- Cover: subject-verb agreement, pronouns, verb tense, parallel structure, modifiers, commas, semicolons, apostrophes, sentence boundaries\n- \"NO CHANGE\" (choice A) correct ~25% of the time, NOT always\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"standard_english_conventions\",\"passage\":\"...with [underlined: ...]...\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[transitions]\n- Passage with ______ blank connecting two ideas. All 4 choices grammatically correct; distinction is logical.\n- Cover: contrast, addition, cause/effect, example, sequence, concession\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"transitions\",\"passage\":\"...______ ...\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[rhetorical_synthesis]\n- Notes format: \"While researching [topic], a student has taken these notes:\\n- Note 1\\n- Note 2\\n- Note 3\"\n- Specific rhetorical goal stated. All 4 choices are grammatical sentences using the notes; only 1 achieves the goal.\n- Goals: emphasize difference, highlight finding, introduce topic, summarize conclusion, present cause-effect\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"rhetorical_synthesis\",\"passage\":\"...\",\"prompt\":\"The student wants to [goal]. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\nOutput ONLY a valid JSON array of EXACTLY 4 objects. No markdown. Count them before outputting.", "text": "[{\"id\":\"sat99-rw-m1-q27\",\"number\":27,\"type\":\"transitions\",\"passage\":\"The axolotl, a salamander native to lakes near Mexico City, retains its larval features—including external gills and a finned tail—throughout its adult life, a phenomenon known as neoteny. Most salamanders undergo metamorphosis, shedding these juvenile traits as they mature. The axolotl, ______ remains perpetually youthful in appearance, a trait that has made it a valuable subject for studies of regeneration and development.\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"by contrast,\",\"B\":\"for instance,\",\"C\":\"as a result,\",\"D\":\"in addition,\"},\"answer_format\":\"mcq\",\"correct_answer\":\"A\"},{\"id\":\"sat99-rw-m1-q28\",\"number\":28,\"type\":\"transitions\",\"passage\":\"Archaeologist Sarah Parcak uses satellite imagery to identify buried archaeological sites, detecting subtle variations in soil and vegetation that hint at structures beneath the surface. Traditional excavation methods can take years and require substantial funding to survey even a small region. Parcak's technique, ______ allows researchers to map hundreds of potential sites in a fraction of the time, dramatically accelerating the pace of discovery.\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"nevertheless,\",\"B\":\"likewise,\",\"C\":\"however,\",\"D\":\"for example,\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m1-q29\",\"number\":29,\"type\":\"transitions\",\"passage\":\"Economist Esther Duflo and her colleagues tested whether small financial incentives could increase childhood immunization rates in rural Rajasthan. Villages offering a kilogram of lentils per vaccination saw immunization rates rise from 6 percent to 38 percent. ______ the program proved more cost-effective than conventional outreach, because the higher participation reduced the per-child expense of operating each mobile clinic.\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"Nonetheless,\",\"B\":\"Similarly,\",\"C\":\"Moreover,\",\"D\":\"Instead,\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m1-q30\",\"number\":30,\"type\":\"transitions\",\"passage\":\"In the late 1870s, painter Mary Cassatt was invited by Edgar Degas to exhibit with the Impressionists in Paris, a group then regarded by critics as radical and unpolished. Cassatt's decision to align herself with these artists cost her the patronage of several conservative American collectors. ______ the affiliation gave her access to a circle of experimental printmakers whose techniques would reshape her own approach to composition.\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"Likewise,\",\"B\":\"In turn,\",\"C\":\"Therefore,\",\"D\":\"Still,\"},\"answer_format\":\"mcq\",\"correct_answer\":\"D\"}]" }, { "costUsd": 0.121365, "model": "opus", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 3 SAT Reading & Writing questions.\nType: rhetorical_synthesis | Test: 99 | Unit: 1 | Start Q#: 31\n\nMANDATORY: Your JSON array MUST contain exactly 3 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nFollow the instructions for your type:\n\n[vocabulary_in_context]\n- 2-3 sentence passage with ______ blank from specific academic domains (molecular biology, Renaissance art, behavioral economics, astrophysics, political philosophy)\n- 4 college-level word choices from the same semantic field\n- Correct word is most PRECISE for context, not most common synonym\n- ID format: sat{testNumber}-rw-m{mod}-q{N}\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"vocabulary_in_context\",\"passage\":\"...with ______...\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[text_comprehension]\n- 3-8 sentence passage from published-quality source (journal, novel, essay) with specific names/dates/findings\n- Mix subtypes: main idea (2-3), author's purpose (1-2), inference (2-3), tone (0-1), structure (1-2)\n- Topics: natural science, social science, literature, history, arts/humanities (use each)\n- Distractors: surface-level misreading, wrong scope, opposite interpretation\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"text_comprehension\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[data_analysis]\n- Passage describing realistic data/study results requiring INTERPRETATION, not lookup\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Data patterns: diminishing returns, optimal values, conditional relationships\n- Distractors: wrong column, wrong trend direction, correlation vs causation confusion\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"data_analysis\",\"passage\":\"...\",\"visual\":{...},\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[claim_support]\n- Mix: \"Which finding supports/undermines the claim?\" AND \"Which quotation illustrates the claim?\"\n- Claims must be specific and testable. Distractors are topically related but don't address the specific mechanism.\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"claim_support\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[standard_english_conventions]\n- Passage with [underlined: text here] portion. Student picks best replacement.\n- Cover: subject-verb agreement, pronouns, verb tense, parallel structure, modifiers, commas, semicolons, apostrophes, sentence boundaries\n- \"NO CHANGE\" (choice A) correct ~25% of the time, NOT always\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"standard_english_conventions\",\"passage\":\"...with [underlined: ...]...\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[transitions]\n- Passage with ______ blank connecting two ideas. All 4 choices grammatically correct; distinction is logical.\n- Cover: contrast, addition, cause/effect, example, sequence, concession\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"transitions\",\"passage\":\"...______ ...\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[rhetorical_synthesis]\n- Notes format: \"While researching [topic], a student has taken these notes:\\n- Note 1\\n- Note 2\\n- Note 3\"\n- Specific rhetorical goal stated. All 4 choices are grammatical sentences using the notes; only 1 achieves the goal.\n- Goals: emphasize difference, highlight finding, introduce topic, summarize conclusion, present cause-effect\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"rhetorical_synthesis\",\"passage\":\"...\",\"prompt\":\"The student wants to [goal]. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\nOutput ONLY a valid JSON array of EXACTLY 3 objects. No markdown. Count them before outputting.", "text": "[{\"id\":\"sat99-rw-m1-q31\",\"number\":31,\"type\":\"rhetorical_synthesis\",\"passage\":\"While researching the migratory patterns of the Arctic tern, a student has taken these notes:\\n- The Arctic tern (Sterna paradisaea) is a small seabird with a wingspan of about 75 cm.\\n- It breeds in Arctic regions during the Northern Hemisphere summer.\\n- It winters in Antarctic waters during the Southern Hemisphere summer.\\n- A 2010 tracking study by Egevang et al. found terns travel roughly 71,000 km annually.\\n- This is the longest known migration of any animal on Earth.\",\"prompt\":\"The student wants to emphasize the remarkable nature of the Arctic tern's migration distance. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"The Arctic tern, a small seabird with a 75 cm wingspan, breeds in the Arctic and winters in Antarctic waters.\",\"B\":\"A 2010 tracking study by Egevang et al. revealed that the Arctic tern travels roughly 71,000 km annually—the longest known migration of any animal on Earth.\",\"C\":\"The Arctic tern breeds in Arctic regions during the Northern Hemisphere summer and migrates to Antarctic waters for the Southern Hemisphere summer.\",\"D\":\"Egevang and colleagues conducted a tracking study of the Arctic tern, a seabird whose scientific name is Sterna paradisaea, in 2010.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m1-q32\",\"number\":32,\"type\":\"rhetorical_synthesis\",\"passage\":\"While researching the painter Artemisia Gentileschi, a student has taken these notes:\\n- Artemisia Gentileschi (1593–1656) was an Italian Baroque painter.\\n- She trained in her father Orazio's workshop in Rome.\\n- In 1616, she became the first woman admitted to the Accademia delle Arti del Disegno in Florence.\\n- Her painting Judith Slaying Holofernes (c. 1620) is held by the Uffizi Gallery.\\n- Art historians cite the work for its dramatic chiaroscuro and unflinching depiction of violence.\",\"prompt\":\"The student wants to introduce Artemisia Gentileschi to an audience unfamiliar with her work. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"Judith Slaying Holofernes, painted around 1620, is noted by art historians for its dramatic chiaroscuro and unflinching depiction of violence.\",\"B\":\"In 1616, Artemisia Gentileschi became the first woman admitted to the Accademia delle Arti del Disegno in Florence.\",\"C\":\"Artemisia Gentileschi (1593–1656) was an Italian Baroque painter whose works, such as Judith Slaying Holofernes, are celebrated for their dramatic intensity.\",\"D\":\"Trained in her father Orazio's Roman workshop, Artemisia Gentileschi later moved to Florence, where the Uffizi Gallery now holds one of her paintings.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m1-q33\",\"number\":33,\"type\":\"rhetorical_synthesis\",\"passage\":\"While researching a behavioral economics experiment, a student has taken these notes:\\n- Researchers Dan Ariely and colleagues studied how anchoring affects willingness to pay.\\n- Participants were asked to write down the last two digits of their Social Security number.\\n- They were then asked to bid on items such as wine, chocolates, and electronics.\\n- Participants with higher digit pairs (80–99) bid up to 346% more than those with lower pairs (00–19).\\n- The digits had no logical connection to the items' actual value.\",\"prompt\":\"The student wants to highlight the central finding of the experiment. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"Dan Ariely and his colleagues conducted a behavioral economics experiment in which participants bid on items including wine, chocolates, and electronics.\",\"B\":\"Although the last two digits of a Social Security number bear no logical relation to an item's value, participants with higher digit pairs bid up to 346% more than those with lower pairs.\",\"C\":\"In one behavioral economics study, participants were asked to write down the last two digits of their Social Security number before bidding on a series of consumer items.\",\"D\":\"Anchoring is a cognitive bias in which an arbitrary number influences subsequent judgments, as Dan Ariely and colleagues demonstrated in their experiment.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"}]" }, { "costUsd": 0.08304, "model": "opus", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 5 SAT Reading & Writing questions.\nType: vocabulary_in_context | Test: 99 | Unit: 2 | Start Q#: 1\n\nMANDATORY: Your JSON array MUST contain exactly 5 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nFollow the instructions for your type:\n\n[vocabulary_in_context]\n- 2-3 sentence passage with ______ blank from specific academic domains (molecular biology, Renaissance art, behavioral economics, astrophysics, political philosophy)\n- 4 college-level word choices from the same semantic field\n- Correct word is most PRECISE for context, not most common synonym\n- ID format: sat{testNumber}-rw-m{mod}-q{N}\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"vocabulary_in_context\",\"passage\":\"...with ______...\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[text_comprehension]\n- 3-8 sentence passage from published-quality source (journal, novel, essay) with specific names/dates/findings\n- Mix subtypes: main idea (2-3), author's purpose (1-2), inference (2-3), tone (0-1), structure (1-2)\n- Topics: natural science, social science, literature, history, arts/humanities (use each)\n- Distractors: surface-level misreading, wrong scope, opposite interpretation\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"text_comprehension\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[data_analysis]\n- Passage describing realistic data/study results requiring INTERPRETATION, not lookup\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Data patterns: diminishing returns, optimal values, conditional relationships\n- Distractors: wrong column, wrong trend direction, correlation vs causation confusion\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"data_analysis\",\"passage\":\"...\",\"visual\":{...},\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[claim_support]\n- Mix: \"Which finding supports/undermines the claim?\" AND \"Which quotation illustrates the claim?\"\n- Claims must be specific and testable. Distractors are topically related but don't address the specific mechanism.\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"claim_support\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[standard_english_conventions]\n- Passage with [underlined: text here] portion. Student picks best replacement.\n- Cover: subject-verb agreement, pronouns, verb tense, parallel structure, modifiers, commas, semicolons, apostrophes, sentence boundaries\n- \"NO CHANGE\" (choice A) correct ~25% of the time, NOT always\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"standard_english_conventions\",\"passage\":\"...with [underlined: ...]...\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[transitions]\n- Passage with ______ blank connecting two ideas. All 4 choices grammatically correct; distinction is logical.\n- Cover: contrast, addition, cause/effect, example, sequence, concession\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"transitions\",\"passage\":\"...______ ...\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[rhetorical_synthesis]\n- Notes format: \"While researching [topic], a student has taken these notes:\\n- Note 1\\n- Note 2\\n- Note 3\"\n- Specific rhetorical goal stated. All 4 choices are grammatical sentences using the notes; only 1 achieves the goal.\n- Goals: emphasize difference, highlight finding, introduce topic, summarize conclusion, present cause-effect\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"rhetorical_synthesis\",\"passage\":\"...\",\"prompt\":\"The student wants to [goal]. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\nOutput ONLY a valid JSON array of EXACTLY 5 objects. No markdown. Count them before outputting.", "text": "[{\"id\":\"sat99-rw-m2-q1\",\"number\":1,\"type\":\"vocabulary_in_context\",\"passage\":\"In studies of chromatin remodeling, researchers have observed that certain ATP-dependent complexes do not merely displace nucleosomes but ______ them along the DNA strand, repositioning histone octamers with remarkable spatial precision.\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"translocate\",\"B\":\"transport\",\"C\":\"transfer\",\"D\":\"transmit\"},\"answer_format\":\"mcq\",\"correct_answer\":\"A\"},{\"id\":\"sat99-rw-m2-q2\",\"number\":2,\"type\":\"vocabulary_in_context\",\"passage\":\"Art historian Patricia Rubin has argued that Vasari's biographies of Renaissance painters were never intended as neutral records; rather, they ______ a Florentine-centered narrative in which artists outside the Medici orbit were systematically diminished. This editorial shaping, she contends, distorted the historiography of Italian art for centuries.\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"disclosed\",\"B\":\"advanced\",\"C\":\"tolerated\",\"D\":\"inherited\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m2-q3\",\"number\":3,\"type\":\"vocabulary_in_context\",\"passage\":\"In behavioral economics, Richard Thaler's concept of mental accounting suggests that consumers do not treat money as fully ______. A tax refund, a birthday gift, and a paycheck of equal value will be spent quite differently, even though rational choice theory predicts identical treatment.\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"divisible\",\"B\":\"transparent\",\"C\":\"fungible\",\"D\":\"negotiable\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m2-q4\",\"number\":4,\"type\":\"vocabulary_in_context\",\"passage\":\"Observations from the James Webb Space Telescope have revealed galaxies at redshifts exceeding $z=13$ that appear far more luminous and structurally mature than existing cosmological models ______. Theorists are now reconsidering whether early star formation proceeded at rates previously thought impossible.\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"accommodate\",\"B\":\"recognize\",\"C\":\"endorse\",\"D\":\"withstand\"},\"answer_format\":\"mcq\",\"correct_answer\":\"A\"},{\"id\":\"sat99-rw-m2-q5\",\"number\":5,\"type\":\"vocabulary_in_context\",\"passage\":\"In \\\"On Liberty,\\\" John Stuart Mill warns that a citizenry may grow so accustomed to the tyranny of majority opinion that dissent becomes not merely unwelcome but nearly ______. Social pressure, he argued, can silence heterodox thought more effectively than any formal censor.\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"unthinkable\",\"B\":\"indifferent\",\"C\":\"redundant\",\"D\":\"ambivalent\"},\"answer_format\":\"mcq\",\"correct_answer\":\"A\"}]" }, { "costUsd": 0.272115, "model": "opus", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 8 SAT Reading & Writing questions.\nType: text_comprehension | Test: 99 | Unit: 2 | Start Q#: 6\n\nMANDATORY: Your JSON array MUST contain exactly 8 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nFollow the instructions for your type:\n\n[vocabulary_in_context]\n- 2-3 sentence passage with ______ blank from specific academic domains (molecular biology, Renaissance art, behavioral economics, astrophysics, political philosophy)\n- 4 college-level word choices from the same semantic field\n- Correct word is most PRECISE for context, not most common synonym\n- ID format: sat{testNumber}-rw-m{mod}-q{N}\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"vocabulary_in_context\",\"passage\":\"...with ______...\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[text_comprehension]\n- 3-8 sentence passage from published-quality source (journal, novel, essay) with specific names/dates/findings\n- Mix subtypes: main idea (2-3), author's purpose (1-2), inference (2-3), tone (0-1), structure (1-2)\n- Topics: natural science, social science, literature, history, arts/humanities (use each)\n- Distractors: surface-level misreading, wrong scope, opposite interpretation\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"text_comprehension\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[data_analysis]\n- Passage describing realistic data/study results requiring INTERPRETATION, not lookup\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Data patterns: diminishing returns, optimal values, conditional relationships\n- Distractors: wrong column, wrong trend direction, correlation vs causation confusion\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"data_analysis\",\"passage\":\"...\",\"visual\":{...},\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[claim_support]\n- Mix: \"Which finding supports/undermines the claim?\" AND \"Which quotation illustrates the claim?\"\n- Claims must be specific and testable. Distractors are topically related but don't address the specific mechanism.\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"claim_support\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[standard_english_conventions]\n- Passage with [underlined: text here] portion. Student picks best replacement.\n- Cover: subject-verb agreement, pronouns, verb tense, parallel structure, modifiers, commas, semicolons, apostrophes, sentence boundaries\n- \"NO CHANGE\" (choice A) correct ~25% of the time, NOT always\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"standard_english_conventions\",\"passage\":\"...with [underlined: ...]...\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[transitions]\n- Passage with ______ blank connecting two ideas. All 4 choices grammatically correct; distinction is logical.\n- Cover: contrast, addition, cause/effect, example, sequence, concession\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"transitions\",\"passage\":\"...______ ...\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[rhetorical_synthesis]\n- Notes format: \"While researching [topic], a student has taken these notes:\\n- Note 1\\n- Note 2\\n- Note 3\"\n- Specific rhetorical goal stated. All 4 choices are grammatical sentences using the notes; only 1 achieves the goal.\n- Goals: emphasize difference, highlight finding, introduce topic, summarize conclusion, present cause-effect\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"rhetorical_synthesis\",\"passage\":\"...\",\"prompt\":\"The student wants to [goal]. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\nOutput ONLY a valid JSON array of EXACTLY 8 objects. No markdown. Count them before outputting.", "text": "[{\"id\":\"sat99-rw-m2-q6\",\"number\":6,\"type\":\"text_comprehension\",\"passage\":\"In her 2019 study of medieval manuscript illumination, art historian Elina Gertsman challenges the long-standing assumption that blank spaces in devotional texts were merely unfinished work awaiting decoration. Drawing on a corpus of over two hundred Books of Hours, Gertsman argues that voids—empty medallions, unpainted margins, and unfilled initials—were often deliberate theological statements. The absence of image, she contends, invited contemplative engagement with the unrepresentable nature of the divine, functioning as a visual corollary to apophatic theology.\",\"prompt\":\"Which choice best describes the main idea of the text?\",\"choices\":{\"A\":\"Gertsman's analysis of Books of Hours reveals that medieval illuminators frequently left manuscripts unfinished due to economic pressures from wealthy patrons.\",\"B\":\"Gertsman proposes that empty spaces in medieval devotional manuscripts should be understood as intentional theological expressions rather than incomplete craftsmanship.\",\"C\":\"Gertsman demonstrates that apophatic theology originated in the artistic practices of medieval manuscript illuminators rather than in formal philosophical treatises.\",\"D\":\"Gertsman's corpus of two hundred Books of Hours proves that medieval viewers preferred unadorned manuscripts to heavily illuminated ones.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m2-q7\",\"number\":7,\"type\":\"text_comprehension\",\"passage\":\"The following is adapted from Willa Cather's 1918 novel My Ántonia. The narrator, Jim Burden, recalls his arrival on the Nebraska prairie as a child.\\n\\nI had the feeling that the world was left behind, that we had got over the edge of it, and were outside man's jurisdiction. I had never before looked up at the sky when there was not a familiar mountain ridge against it. But this was the complete dome of heaven, all there was of it. Between that earth and that sky I felt erased, blotted out. I did not say my prayers that night: here, I felt, what would be would be.\",\"prompt\":\"Which choice best describes the function of the underlined sentence \\\"Between that earth and that sky I felt erased, blotted out\\\" in the text as a whole?\",\"choices\":{\"A\":\"It introduces a geographic description that the narrator will later contradict with more detailed observations.\",\"B\":\"It conveys the narrator's fear that the prairie's inhabitants will not welcome a newcomer from the East.\",\"C\":\"It captures a sense of self-dissolution that motivates the narrator's subsequent abandonment of a familiar ritual.\",\"D\":\"It criticizes the narrator's earlier assumption that mountains were necessary for a meaningful landscape.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m2-q8\",\"number\":8,\"type\":\"text_comprehension\",\"passage\":\"Economist Esther Duflo's randomized controlled trials in rural Rajasthan produced a counterintuitive finding: offering mothers a small bag of lentils as an incentive for vaccinating their children raised full-immunization rates from 6 percent to 39 percent, while the program's per-child cost actually fell. The lentils, Duflo observed, did not persuade skeptics of vaccine safety; rather, they solved a subtler problem. Many parents already intended to immunize but perpetually postponed the clinic visit. The modest reward supplied the nudge needed to convert latent intention into action.\",\"prompt\":\"Based on the text, what can reasonably be inferred about the parents in Duflo's study?\",\"choices\":{\"A\":\"Most parents were skeptical of vaccine safety and required material incentives to overcome their doubts.\",\"B\":\"Parents in rural Rajasthan generally placed greater value on food staples than on their children's long-term health.\",\"C\":\"Many parents' failure to vaccinate reflected procrastination rather than an objection to immunization itself.\",\"D\":\"The program's success depended primarily on reducing the travel distance between villages and immunization clinics.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m2-q9\",\"number\":9,\"type\":\"text_comprehension\",\"passage\":\"Paleontologist Jingmai O'Connor has argued that the traditional dichotomy between \\\"avian\\\" and \\\"non-avian\\\" dinosaurs obscures more than it clarifies. Her analyses of Cretaceous specimens from the Jehol Biota suggest that features once considered diagnostic of birds—feathers, wishbones, even rudimentary flight—appeared sporadically across multiple theropod lineages. To speak of \\\"the origin of birds,\\\" O'Connor writes, is to impose a retrospective clarity on what was, in the Mesozoic, a thicket of parallel experiments.\",\"prompt\":\"Which choice best describes the author's purpose in the text?\",\"choices\":{\"A\":\"To argue that O'Connor's research has definitively resolved longstanding debates about which theropod lineage gave rise to modern birds.\",\"B\":\"To explain how O'Connor's work complicates a familiar evolutionary narrative by emphasizing the distributed nature of bird-like traits.\",\"C\":\"To criticize paleontologists who continue to use the terms \\\"avian\\\" and \\\"non-avian\\\" despite evidence from the Jehol Biota.\",\"D\":\"To describe a methodological controversy about how fossil specimens from the Cretaceous should be classified and catalogued.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m2-q10\",\"number\":10,\"type\":\"text_comprehension\",\"passage\":\"In a 2021 review of urban heat-island mitigation, civil engineer Hashem Akbari surveyed three decades of rooftop albedo studies. He found that while \\\"cool roofs\\\" painted with highly reflective coatings reliably reduced indoor cooling loads by 10–20 percent in low-rise buildings, their neighborhood-scale effects were far less consistent. In some Mediterranean cities, reflected solar radiation from dense clusters of cool roofs modestly raised street-level temperatures for pedestrians, particularly in narrow canyons where the reflected light was repeatedly absorbed by adjacent walls.\",\"prompt\":\"Which choice best describes the structure of the text?\",\"choices\":{\"A\":\"It presents a widely accepted engineering solution and then describes the mechanism by which it was originally developed.\",\"B\":\"It contrasts two competing theories about urban heat mitigation and argues in favor of the more recent one.\",\"C\":\"It introduces a well-established benefit of a technology and then qualifies that benefit by noting a context-dependent drawback.\",\"D\":\"It summarizes a historical debate and concludes that the debate has been resolved by Akbari's 2021 review.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m2-q11\",\"number\":11,\"type\":\"text_comprehension\",\"passage\":\"The following is adapted from an 1891 essay by Vernon Lee on the aesthetics of music.\\n\\nI do not deny that a sonata may be dissected, its modulations charted, its thematic transformations labeled with scholarly precision. Such labor has its uses. But let us not pretend that the sum of these annotations is the sonata. The listener who emerges from the concert hall transfigured has not been moved by a catalogue of devices; she has been moved by something the catalogue cannot capture, and which the catalogue was never designed to record.\",\"prompt\":\"Which choice best describes the tone of the passage?\",\"choices\":{\"A\":\"Dismissive—Lee rejects musical analysis as a futile endeavor that produces no genuine understanding.\",\"B\":\"Measured—Lee concedes the value of technical analysis while insisting on its insufficiency for explaining musical experience.\",\"C\":\"Nostalgic—Lee laments the decline of an earlier era in which listeners were more emotionally responsive to music.\",\"D\":\"Defensive—Lee justifies her own lack of formal musical training by attacking those who possess it.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m2-q12\",\"number\":12,\"type\":\"text_comprehension\",\"passage\":\"Historians of the early Qing dynasty have long debated the significance of the 1645 queue edict, which required Han Chinese men to adopt the Manchu hairstyle on pain of death. Recent work by Joanna Waley-Cohen reframes the edict as less an assertion of ethnic supremacy than a calculated test of political loyalty. Because the hairstyle was visible at a glance, it allowed Qing officials to identify dissent without the laborious inquisitions that had undermined earlier dynasties. The edict's cruelty, Waley-Cohen suggests, should not obscure its administrative elegance.\",\"prompt\":\"Based on the text, which of the following best describes Waley-Cohen's interpretation of the queue edict?\",\"choices\":{\"A\":\"It was primarily an expression of Manchu cultural chauvinism against Han Chinese traditions.\",\"B\":\"It was a comparatively humane policy when measured against the inquisitions of earlier dynasties.\",\"C\":\"It functioned as an efficient surveillance mechanism that reduced the need for more invasive forms of investigation.\",\"D\":\"It was an impractical policy whose administrative costs outweighed its benefits to the Qing state.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m2-q13\",\"number\":13,\"type\":\"text_comprehension\",\"passage\":\"When the astronomer Vera Rubin began measuring the rotation curves of spiral galaxies in the 1970s, she expected the outer stars to orbit more slowly than those near the galactic center, just as distant planets orbit the sun more slowly than nearby ones. Instead, her spectra showed that stars at the edges of galaxies like Andromeda were moving just as fast as those much closer in. The implication, Rubin recognized, was either that the law of gravity behaved differently on galactic scales or that each galaxy was embedded in a vast halo of unseen mass.\",\"prompt\":\"Which choice best states the main idea of the text?\",\"choices\":{\"A\":\"Rubin's spectra of Andromeda disproved the Newtonian laws of gravity and established a new framework for galactic dynamics.\",\"B\":\"Rubin's observations of galactic rotation contradicted her expectations and pointed toward two possible explanations with profound implications.\",\"C\":\"Rubin's work demonstrated that spiral galaxies rotate differently from solar systems because their stars are more massive than planets.\",\"D\":\"Rubin initially misinterpreted her spectra but later corrected her findings after comparing them with planetary orbital data.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"}]" }, { "costUsd": 0.17019, "model": "opus", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 4 SAT Reading & Writing questions.\nType: data_analysis | Test: 99 | Unit: 2 | Start Q#: 14\n\nMANDATORY: Your JSON array MUST contain exactly 4 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nFollow the instructions for your type:\n\n[vocabulary_in_context]\n- 2-3 sentence passage with ______ blank from specific academic domains (molecular biology, Renaissance art, behavioral economics, astrophysics, political philosophy)\n- 4 college-level word choices from the same semantic field\n- Correct word is most PRECISE for context, not most common synonym\n- ID format: sat{testNumber}-rw-m{mod}-q{N}\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"vocabulary_in_context\",\"passage\":\"...with ______...\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[text_comprehension]\n- 3-8 sentence passage from published-quality source (journal, novel, essay) with specific names/dates/findings\n- Mix subtypes: main idea (2-3), author's purpose (1-2), inference (2-3), tone (0-1), structure (1-2)\n- Topics: natural science, social science, literature, history, arts/humanities (use each)\n- Distractors: surface-level misreading, wrong scope, opposite interpretation\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"text_comprehension\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[data_analysis]\n- Passage describing realistic data/study results requiring INTERPRETATION, not lookup\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Data patterns: diminishing returns, optimal values, conditional relationships\n- Distractors: wrong column, wrong trend direction, correlation vs causation confusion\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"data_analysis\",\"passage\":\"...\",\"visual\":{...},\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[claim_support]\n- Mix: \"Which finding supports/undermines the claim?\" AND \"Which quotation illustrates the claim?\"\n- Claims must be specific and testable. Distractors are topically related but don't address the specific mechanism.\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"claim_support\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[standard_english_conventions]\n- Passage with [underlined: text here] portion. Student picks best replacement.\n- Cover: subject-verb agreement, pronouns, verb tense, parallel structure, modifiers, commas, semicolons, apostrophes, sentence boundaries\n- \"NO CHANGE\" (choice A) correct ~25% of the time, NOT always\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"standard_english_conventions\",\"passage\":\"...with [underlined: ...]...\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[transitions]\n- Passage with ______ blank connecting two ideas. All 4 choices grammatically correct; distinction is logical.\n- Cover: contrast, addition, cause/effect, example, sequence, concession\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"transitions\",\"passage\":\"...______ ...\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[rhetorical_synthesis]\n- Notes format: \"While researching [topic], a student has taken these notes:\\n- Note 1\\n- Note 2\\n- Note 3\"\n- Specific rhetorical goal stated. All 4 choices are grammatical sentences using the notes; only 1 achieves the goal.\n- Goals: emphasize difference, highlight finding, introduce topic, summarize conclusion, present cause-effect\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"rhetorical_synthesis\",\"passage\":\"...\",\"prompt\":\"The student wants to [goal]. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\nOutput ONLY a valid JSON array of EXACTLY 4 objects. No markdown. Count them before outputting.", "text": "[{\"id\":\"sat99-rw-m2-q14\",\"number\":14,\"type\":\"data_analysis\",\"passage\":\"Soil ecologist Marisol Vega investigated whether mycorrhizal fungi inoculation affects the drought tolerance of seedlings in degraded pastureland. She grew seedlings of four native tree species under identical water-restricted conditions, half with inoculation and half without, and measured survival rates after 90 days. Vega hypothesized that the benefit of inoculation would be greatest for species with naturally low drought tolerance, as measured by baseline (non-inoculated) survival.\",\"visual\":{\"type\":\"table\",\"title\":\"90-Day Seedling Survival Rates Under Drought Stress (%)\",\"headers\":[\"Species\",\"Non-inoculated\",\"Inoculated\",\"Change (pp)\"],\"rows\":[[\"Quercus rugosa\",\"72\",\"81\",\"+9\"],[\"Pinus leiophylla\",\"58\",\"74\",\"+16\"],[\"Alnus acuminata\",\"34\",\"62\",\"+28\"],[\"Prunus serotina\",\"49\",\"65\",\"+16\"]]},\"prompt\":\"Which choice best describes data from the table that support Vega's hypothesis?\",\"choices\":{\"A\":\"Alnus acuminata had the lowest non-inoculated survival rate and showed the largest increase in survival following inoculation.\",\"B\":\"Quercus rugosa had the highest non-inoculated survival rate and showed the largest increase in survival following inoculation.\",\"C\":\"Pinus leiophylla and Prunus serotina had identical increases in survival despite different baseline survival rates.\",\"D\":\"Every species showed an increase in survival following inoculation, with the average increase exceeding 15 percentage points.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"A\"},{\"id\":\"sat99-rw-m2-q15\",\"number\":15,\"type\":\"data_analysis\",\"passage\":\"In a study of consumer decision-making, economist Tarun Shah presented participants with a choice between two retirement savings plans that differed only in how contribution rates were framed: Plan X described contributions as a percentage of gross salary, while Plan Y described the same contributions as a fixed dollar amount per paycheck. Shah recorded the proportion of participants in each age group who selected Plan Y.\",\"visual\":{\"type\":\"table\",\"title\":\"Proportion Selecting Plan Y (Fixed-Dollar Framing) by Age Group\",\"headers\":[\"Age Group\",\"Plan Y Selection Rate\",\"Reported Confidence in Choice\"],\"rows\":[[\"18–29\",\"0.71\",\"Low\"],[\"30–44\",\"0.58\",\"Moderate\"],[\"45–59\",\"0.44\",\"Moderate\"],[\"60+\",\"0.29\",\"High\"]]},\"prompt\":\"Based on the table, which statement most accurately characterizes the relationship between age and framing preference?\",\"choices\":{\"A\":\"Older participants were more likely to select the fixed-dollar plan, suggesting increased susceptibility to concrete framing with age.\",\"B\":\"As age increased, preference for the fixed-dollar framing decreased, and reported confidence in the choice increased.\",\"C\":\"Confidence in the choice was lowest among participants who most frequently selected the percentage-based framing.\",\"D\":\"Age had no systematic effect on framing preference, since selection rates varied unpredictably across groups.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m2-q16\",\"number\":16,\"type\":\"data_analysis\",\"passage\":\"To assess whether urban noise affects the vocal behavior of songbirds, ornithologist Kenji Arai recorded dawn choruses of white-crowned sparrows at four sites in and around San Francisco. At each site, he measured the ambient noise level (in decibels) at 5:00 a.m. and the minimum frequency (in hertz) of the sparrows' songs, reasoning that birds in noisier environments might raise their minimum song frequency to avoid masking by low-frequency traffic noise.\",\"visual\":{\"type\":\"table\",\"title\":\"Ambient Noise and Minimum Song Frequency at Four Sites\",\"headers\":[\"Site\",\"Ambient Noise (dB)\",\"Minimum Song Frequency (Hz)\"],\"rows\":[[\"Presidio (suburban edge)\",\"42\",\"2,050\"],[\"Golden Gate Park\",\"51\",\"2,310\"],[\"Mission District\",\"63\",\"2,740\"],[\"SOMA (downtown)\",\"71\",\"3,020\"]]},\"prompt\":\"Which choice best describes data from the table that support Arai's reasoning?\",\"choices\":{\"A\":\"The Presidio site had the lowest ambient noise and the lowest minimum song frequency, while SOMA had the highest values for both.\",\"B\":\"Minimum song frequency at each site was more than 40 times the ambient noise level measured at that site.\",\"C\":\"The difference in minimum song frequency between Golden Gate Park and the Mission District was greater than the difference between the Presidio and Golden Gate Park.\",\"D\":\"Sites with higher ambient noise showed lower minimum song frequencies, contrary to what would be expected if traffic noise masked low-frequency song.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"A\"},{\"id\":\"sat99-rw-m2-q17\",\"number\":17,\"type\":\"data_analysis\",\"passage\":\"Historian Elena Costa examined the circulation numbers of four political pamphlets published in Venice between 1575 and 1590 to investigate whether length influenced popular reach. She hypothesized that shorter pamphlets would achieve wider circulation because they were cheaper to print and easier for semi-literate readers to consume in a single sitting.\",\"visual\":{\"type\":\"table\",\"title\":\"Venetian Political Pamphlets, 1575–1590\",\"headers\":[\"Pamphlet\",\"Length (pages)\",\"Estimated Copies Circulated\",\"Price (soldi)\"],\"rows\":[[\"Discorso sulla Repubblica\",\"48\",\"1,200\",\"12\"],[\"Lettera ai Cittadini\",\"8\",\"9,400\",\"2\"],[\"Trattato della Giustizia\",\"64\",\"850\",\"16\"],[\"Avviso al Popolo\",\"12\",\"7,100\",\"3\"]]},\"prompt\":\"Which choice best describes data from the table that support Costa's hypothesis?\",\"choices\":{\"A\":\"The two longest pamphlets, each exceeding 40 pages, circulated in fewer than 1,500 copies, while the two shortest, under 15 pages, each circulated in more than 7,000 copies.\",\"B\":\"Lettera ai Cittadini circulated in more copies than any other pamphlet, and its price was the lowest of the four.\",\"C\":\"The pamphlet with the highest price, Trattato della Giustizia, circulated in the fewest copies, suggesting that cost rather than length drove readership.\",\"D\":\"Avviso al Popolo was longer than Lettera ai Cittadini but circulated in fewer copies, showing that length had only a modest effect on reach.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"A\"}]" }, { "costUsd": 0.145665, "model": "opus", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 4 SAT Reading & Writing questions.\nType: claim_support | Test: 99 | Unit: 2 | Start Q#: 18\n\nMANDATORY: Your JSON array MUST contain exactly 4 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nFollow the instructions for your type:\n\n[vocabulary_in_context]\n- 2-3 sentence passage with ______ blank from specific academic domains (molecular biology, Renaissance art, behavioral economics, astrophysics, political philosophy)\n- 4 college-level word choices from the same semantic field\n- Correct word is most PRECISE for context, not most common synonym\n- ID format: sat{testNumber}-rw-m{mod}-q{N}\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"vocabulary_in_context\",\"passage\":\"...with ______...\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[text_comprehension]\n- 3-8 sentence passage from published-quality source (journal, novel, essay) with specific names/dates/findings\n- Mix subtypes: main idea (2-3), author's purpose (1-2), inference (2-3), tone (0-1), structure (1-2)\n- Topics: natural science, social science, literature, history, arts/humanities (use each)\n- Distractors: surface-level misreading, wrong scope, opposite interpretation\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"text_comprehension\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[data_analysis]\n- Passage describing realistic data/study results requiring INTERPRETATION, not lookup\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Data patterns: diminishing returns, optimal values, conditional relationships\n- Distractors: wrong column, wrong trend direction, correlation vs causation confusion\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"data_analysis\",\"passage\":\"...\",\"visual\":{...},\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[claim_support]\n- Mix: \"Which finding supports/undermines the claim?\" AND \"Which quotation illustrates the claim?\"\n- Claims must be specific and testable. Distractors are topically related but don't address the specific mechanism.\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"claim_support\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[standard_english_conventions]\n- Passage with [underlined: text here] portion. Student picks best replacement.\n- Cover: subject-verb agreement, pronouns, verb tense, parallel structure, modifiers, commas, semicolons, apostrophes, sentence boundaries\n- \"NO CHANGE\" (choice A) correct ~25% of the time, NOT always\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"standard_english_conventions\",\"passage\":\"...with [underlined: ...]...\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[transitions]\n- Passage with ______ blank connecting two ideas. All 4 choices grammatically correct; distinction is logical.\n- Cover: contrast, addition, cause/effect, example, sequence, concession\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"transitions\",\"passage\":\"...______ ...\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[rhetorical_synthesis]\n- Notes format: \"While researching [topic], a student has taken these notes:\\n- Note 1\\n- Note 2\\n- Note 3\"\n- Specific rhetorical goal stated. All 4 choices are grammatical sentences using the notes; only 1 achieves the goal.\n- Goals: emphasize difference, highlight finding, introduce topic, summarize conclusion, present cause-effect\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"rhetorical_synthesis\",\"passage\":\"...\",\"prompt\":\"The student wants to [goal]. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\nOutput ONLY a valid JSON array of EXACTLY 4 objects. No markdown. Count them before outputting.", "text": "[{\"id\":\"sat99-rw-m2-q18\",\"number\":18,\"type\":\"claim_support\",\"passage\":\"In a 2019 study, marine biologist Elena Vasquez examined the foraging behavior of Caribbean reef octopuses (Octopus briareus) and proposed that these cephalopods develop individualized hunting strategies that persist across multiple foraging trips, rather than responding identically to similar environmental stimuli. Vasquez argued that this consistency reflects stable behavioral differences between individuals analogous to personality traits in vertebrates.\",\"prompt\":\"Which finding from Vasquez's study, if true, would most directly support her claim?\",\"choices\":{\"A\":\"Octopuses in the study showed a marked preference for crustacean prey over mollusks, regardless of which prey type was more abundant in their immediate habitat.\",\"B\":\"When presented with identical artificial reef environments, individual octopuses each used distinct hunting techniques—such as ambush, pursuit, or den-baiting—that each animal reliably repeated across twelve separate trials.\",\"C\":\"Octopuses observed in the wild exhibited more varied hunting behaviors than those kept in controlled laboratory conditions, suggesting that environmental complexity shapes foraging choices.\",\"D\":\"Juvenile octopuses in the study displayed a wider range of hunting techniques than mature adults, whose behaviors became increasingly uniform as they aged.\",\"E\":\"\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m2-q19\",\"number\":19,\"type\":\"claim_support\",\"passage\":\"In her 2021 analysis of Dutch Golden Age still life painting, art historian Margriet de Vries contends that Clara Peeters's signature technique of embedding tiny self-portraits within the reflective surfaces of depicted pewter vessels was not merely a display of technical virtuosity but a deliberate assertion of authorial presence in a genre that typically effaced the painter's identity. De Vries argues that these reflections functioned as visual signatures at a time when female artists were frequently denied formal attribution.\",\"prompt\":\"Which finding, if true, would most strongly support de Vries's claim?\",\"choices\":{\"A\":\"X-ray analysis of several Peeters paintings reveals that the reflected self-portraits were painted in a final, separate layer using pigments distinct from those used in the surrounding still life composition.\",\"B\":\"Peeters's contemporaries, including male still life painters such as Osias Beert, occasionally included reflections of windows or studio interiors in their depicted metalware.\",\"C\":\"Archival records indicate that several of Peeters's still lifes were sold to wealthy patrons in Antwerp without any written attribution to the artist, and that she signed far fewer works than her male contemporaries.\",\"D\":\"Technical examination of pewter vessels from the early seventeenth century shows that their polished surfaces would have produced recognizable mirror-like reflections under studio lighting.\",\"E\":\"\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m2-q20\",\"number\":20,\"type\":\"claim_support\",\"passage\":\"In The Theory of Moral Sentiments (1759), Adam Smith argues that human beings possess an innate capacity for sympathy, which he defines not as pity but as the imaginative act of placing oneself in another's situation. Smith further claims that our moral judgments of others arise primarily from this sympathetic projection rather than from abstract reasoning about duty or consequence.\",\"prompt\":\"Which quotation from The Theory of Moral Sentiments most directly illustrates Smith's claim about the origin of moral judgment?\",\"choices\":{\"A\":\"\\\"Mankind, though naturally sympathetic, never conceive, for what has befallen another, that degree of passion which naturally animates the person principally concerned.\\\"\",\"B\":\"\\\"The man of the most perfect virtue... is he who joins, to the most perfect command of his own original and selfish feelings, the most exquisite sensibility both to the original and sympathetic feelings of others.\\\"\",\"C\":\"\\\"As we have no immediate experience of what other men feel, we can form no idea of the manner in which they are affected, but by conceiving what we ourselves should feel in the like situation.\\\"\",\"D\":\"\\\"Every faculty in one man is the measure by which he judges of the like faculty in another. I judge of your sight by my sight, of your ear by my ear, of your reason by my reason.\\\"\",\"E\":\"\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m2-q21\",\"number\":21,\"type\":\"claim_support\",\"passage\":\"Behavioral economist Priya Narayanan has proposed that so-called \\\"default effects\\\"—the tendency of people to stick with pre-selected options—stem not from genuine preference alignment but from the cognitive effort required to evaluate alternatives. Narayanan predicts that when decision-makers are given more time and cognitive resources to deliberate, the influence of defaults on their final choices will diminish substantially.\",\"prompt\":\"Which finding from a controlled experiment would most directly undermine Narayanan's claim?\",\"choices\":{\"A\":\"Participants who were given a forty-five minute deliberation period and reminded to consider all options chose the default option at essentially the same rate as participants who were required to decide within thirty seconds.\",\"B\":\"Participants under time pressure were more likely to report feeling anxious about their decisions than participants given ample time to deliberate, though both groups selected the default at similar rates.\",\"C\":\"Participants who were required to justify their decisions in writing chose the default less often than those who were not asked to provide justifications, regardless of how much time they were given.\",\"D\":\"Participants who were explicitly told which option had been pre-selected as the default were more likely to choose it than participants who were not informed about the default's status.\",\"E\":\"\"},\"answer_format\":\"mcq\",\"correct_answer\":\"A\"}]" }, { "costUsd": 0.07389, "model": "opus", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 5 SAT Reading & Writing questions.\nType: standard_english_conventions | Test: 99 | Unit: 2 | Start Q#: 22\n\nMANDATORY: Your JSON array MUST contain exactly 5 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nFollow the instructions for your type:\n\n[vocabulary_in_context]\n- 2-3 sentence passage with ______ blank from specific academic domains (molecular biology, Renaissance art, behavioral economics, astrophysics, political philosophy)\n- 4 college-level word choices from the same semantic field\n- Correct word is most PRECISE for context, not most common synonym\n- ID format: sat{testNumber}-rw-m{mod}-q{N}\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"vocabulary_in_context\",\"passage\":\"...with ______...\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[text_comprehension]\n- 3-8 sentence passage from published-quality source (journal, novel, essay) with specific names/dates/findings\n- Mix subtypes: main idea (2-3), author's purpose (1-2), inference (2-3), tone (0-1), structure (1-2)\n- Topics: natural science, social science, literature, history, arts/humanities (use each)\n- Distractors: surface-level misreading, wrong scope, opposite interpretation\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"text_comprehension\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[data_analysis]\n- Passage describing realistic data/study results requiring INTERPRETATION, not lookup\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Data patterns: diminishing returns, optimal values, conditional relationships\n- Distractors: wrong column, wrong trend direction, correlation vs causation confusion\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"data_analysis\",\"passage\":\"...\",\"visual\":{...},\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[claim_support]\n- Mix: \"Which finding supports/undermines the claim?\" AND \"Which quotation illustrates the claim?\"\n- Claims must be specific and testable. Distractors are topically related but don't address the specific mechanism.\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"claim_support\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[standard_english_conventions]\n- Passage with [underlined: text here] portion. Student picks best replacement.\n- Cover: subject-verb agreement, pronouns, verb tense, parallel structure, modifiers, commas, semicolons, apostrophes, sentence boundaries\n- \"NO CHANGE\" (choice A) correct ~25% of the time, NOT always\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"standard_english_conventions\",\"passage\":\"...with [underlined: ...]...\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[transitions]\n- Passage with ______ blank connecting two ideas. All 4 choices grammatically correct; distinction is logical.\n- Cover: contrast, addition, cause/effect, example, sequence, concession\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"transitions\",\"passage\":\"...______ ...\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[rhetorical_synthesis]\n- Notes format: \"While researching [topic], a student has taken these notes:\\n- Note 1\\n- Note 2\\n- Note 3\"\n- Specific rhetorical goal stated. All 4 choices are grammatical sentences using the notes; only 1 achieves the goal.\n- Goals: emphasize difference, highlight finding, introduce topic, summarize conclusion, present cause-effect\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"rhetorical_synthesis\",\"passage\":\"...\",\"prompt\":\"The student wants to [goal]. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\nOutput ONLY a valid JSON array of EXACTLY 5 objects. No markdown. Count them before outputting.", "text": "[{\"id\":\"sat99-rw-m2-q22\",\"number\":22,\"type\":\"standard_english_conventions\",\"passage\":\"The archaeologist Kathleen Kenyon, whose excavations at Jericho in the 1950s revolutionized the field's understanding of Neolithic settlement, [underlined: are] now regarded as a pioneer of stratigraphic methodology.\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"is\",\"C\":\"were\",\"D\":\"being\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m2-q23\",\"number\":23,\"type\":\"standard_english_conventions\",\"passage\":\"Marine biologists studying the Great Barrier Reef have documented a surprising resilience in certain coral species; [underlined: however these] organisms continue to reproduce even after repeated bleaching events.\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"however, these\",\"C\":\"however: these\",\"D\":\"however these,\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m2-q24\",\"number\":24,\"type\":\"standard_english_conventions\",\"passage\":\"When composing her final symphony, Florence Price drew on spirituals, European classical forms, and [underlined: she incorporated syncopated rhythms] characteristic of African American folk traditions.\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"incorporating syncopated rhythms\",\"C\":\"syncopated rhythms\",\"D\":\"syncopated rhythms were incorporated\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m2-q25\",\"number\":25,\"type\":\"standard_english_conventions\",\"passage\":\"The committee's decision to fund the observatory surprised many researchers [underlined: who's] grant proposals had been rejected in earlier rounds of review.\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"whose\",\"C\":\"whos'\",\"D\":\"whom's\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m2-q26\",\"number\":26,\"type\":\"standard_english_conventions\",\"passage\":\"By the time the expedition reached the base camp on the southern glacier, the team [underlined: has endured] three weeks of subzero temperatures and persistent whiteout conditions.\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"endures\",\"C\":\"is enduring\",\"D\":\"had endured\"},\"answer_format\":\"mcq\",\"correct_answer\":\"D\"}]" }, { "costUsd": 0.082365, "model": "opus", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 4 SAT Reading & Writing questions.\nType: transitions | Test: 99 | Unit: 2 | Start Q#: 27\n\nMANDATORY: Your JSON array MUST contain exactly 4 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nFollow the instructions for your type:\n\n[vocabulary_in_context]\n- 2-3 sentence passage with ______ blank from specific academic domains (molecular biology, Renaissance art, behavioral economics, astrophysics, political philosophy)\n- 4 college-level word choices from the same semantic field\n- Correct word is most PRECISE for context, not most common synonym\n- ID format: sat{testNumber}-rw-m{mod}-q{N}\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"vocabulary_in_context\",\"passage\":\"...with ______...\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[text_comprehension]\n- 3-8 sentence passage from published-quality source (journal, novel, essay) with specific names/dates/findings\n- Mix subtypes: main idea (2-3), author's purpose (1-2), inference (2-3), tone (0-1), structure (1-2)\n- Topics: natural science, social science, literature, history, arts/humanities (use each)\n- Distractors: surface-level misreading, wrong scope, opposite interpretation\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"text_comprehension\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[data_analysis]\n- Passage describing realistic data/study results requiring INTERPRETATION, not lookup\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Data patterns: diminishing returns, optimal values, conditional relationships\n- Distractors: wrong column, wrong trend direction, correlation vs causation confusion\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"data_analysis\",\"passage\":\"...\",\"visual\":{...},\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[claim_support]\n- Mix: \"Which finding supports/undermines the claim?\" AND \"Which quotation illustrates the claim?\"\n- Claims must be specific and testable. Distractors are topically related but don't address the specific mechanism.\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"claim_support\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[standard_english_conventions]\n- Passage with [underlined: text here] portion. Student picks best replacement.\n- Cover: subject-verb agreement, pronouns, verb tense, parallel structure, modifiers, commas, semicolons, apostrophes, sentence boundaries\n- \"NO CHANGE\" (choice A) correct ~25% of the time, NOT always\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"standard_english_conventions\",\"passage\":\"...with [underlined: ...]...\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[transitions]\n- Passage with ______ blank connecting two ideas. All 4 choices grammatically correct; distinction is logical.\n- Cover: contrast, addition, cause/effect, example, sequence, concession\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"transitions\",\"passage\":\"...______ ...\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[rhetorical_synthesis]\n- Notes format: \"While researching [topic], a student has taken these notes:\\n- Note 1\\n- Note 2\\n- Note 3\"\n- Specific rhetorical goal stated. All 4 choices are grammatical sentences using the notes; only 1 achieves the goal.\n- Goals: emphasize difference, highlight finding, introduce topic, summarize conclusion, present cause-effect\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"rhetorical_synthesis\",\"passage\":\"...\",\"prompt\":\"The student wants to [goal]. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\nOutput ONLY a valid JSON array of EXACTLY 4 objects. No markdown. Count them before outputting.", "text": "[{\"id\":\"sat99-rw-m2-q27\",\"number\":27,\"type\":\"transitions\",\"passage\":\"The axolotl, a salamander native to lakes near Mexico City, retains its larval features—including external gills and a finned tail—throughout adulthood, a phenomenon known as neoteny. Most amphibians undergo metamorphosis to develop lungs and adapt to terrestrial life. ______ the axolotl remains fully aquatic, breeding and feeding in the same form it had as a juvenile.\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"By contrast,\",\"B\":\"Likewise,\",\"C\":\"As a result,\",\"D\":\"For instance,\"},\"answer_format\":\"mcq\",\"correct_answer\":\"A\"},{\"id\":\"sat99-rw-m2-q28\",\"number\":28,\"type\":\"transitions\",\"passage\":\"In her 2019 study of urban tree canopies, ecologist Mariana Velez found that neighborhoods with greater tree cover recorded summer temperatures up to 4°C lower than adjacent treeless blocks. The cooling effect persisted even during prolonged heat waves, when air-conditioning demand spiked elsewhere. ______ Velez concluded that municipal tree-planting initiatives could meaningfully reduce both energy consumption and heat-related illness in dense city centers.\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"Nevertheless,\",\"B\":\"On these grounds,\",\"C\":\"In addition,\",\"D\":\"Granted,\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m2-q29\",\"number\":29,\"type\":\"transitions\",\"passage\":\"Composer Florence Price submitted her Symphony in E minor to the 1932 Wanamaker Foundation competition and won first prize, prompting the Chicago Symphony Orchestra to perform the work the following year. For decades afterward, however, her scores were largely forgotten and her name absent from concert programs. ______ in 2009, a renovation crew discovered a trove of Price's manuscripts in her abandoned summer home, sparking renewed scholarly and performance interest.\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"Consequently,\",\"B\":\"Similarly,\",\"C\":\"Then,\",\"D\":\"Meanwhile,\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m2-q30\",\"number\":30,\"type\":\"transitions\",\"passage\":\"Behavioral economist Sendhil Mullainathan has argued that scarcity—whether of time, money, or food—captures the mind, narrowing attention to the immediate shortage and degrading performance on unrelated tasks. Critics initially dismissed the claim as overstated, suggesting that observed deficits stemmed from chronic stress rather than scarcity itself. ______ subsequent experiments in which participants were temporarily induced to feel scarce, and then released from that state, replicated the cognitive effects within a single laboratory session.\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"Moreover,\",\"B\":\"For example,\",\"C\":\"In short,\",\"D\":\"However,\"},\"answer_format\":\"mcq\",\"correct_answer\":\"D\"}]" }, { "costUsd": 0.12909, "model": "opus", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 3 SAT Reading & Writing questions.\nType: rhetorical_synthesis | Test: 99 | Unit: 2 | Start Q#: 31\n\nMANDATORY: Your JSON array MUST contain exactly 3 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nFollow the instructions for your type:\n\n[vocabulary_in_context]\n- 2-3 sentence passage with ______ blank from specific academic domains (molecular biology, Renaissance art, behavioral economics, astrophysics, political philosophy)\n- 4 college-level word choices from the same semantic field\n- Correct word is most PRECISE for context, not most common synonym\n- ID format: sat{testNumber}-rw-m{mod}-q{N}\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"vocabulary_in_context\",\"passage\":\"...with ______...\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[text_comprehension]\n- 3-8 sentence passage from published-quality source (journal, novel, essay) with specific names/dates/findings\n- Mix subtypes: main idea (2-3), author's purpose (1-2), inference (2-3), tone (0-1), structure (1-2)\n- Topics: natural science, social science, literature, history, arts/humanities (use each)\n- Distractors: surface-level misreading, wrong scope, opposite interpretation\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"text_comprehension\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[data_analysis]\n- Passage describing realistic data/study results requiring INTERPRETATION, not lookup\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Data patterns: diminishing returns, optimal values, conditional relationships\n- Distractors: wrong column, wrong trend direction, correlation vs causation confusion\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"data_analysis\",\"passage\":\"...\",\"visual\":{...},\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[claim_support]\n- Mix: \"Which finding supports/undermines the claim?\" AND \"Which quotation illustrates the claim?\"\n- Claims must be specific and testable. Distractors are topically related but don't address the specific mechanism.\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"claim_support\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[standard_english_conventions]\n- Passage with [underlined: text here] portion. Student picks best replacement.\n- Cover: subject-verb agreement, pronouns, verb tense, parallel structure, modifiers, commas, semicolons, apostrophes, sentence boundaries\n- \"NO CHANGE\" (choice A) correct ~25% of the time, NOT always\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"standard_english_conventions\",\"passage\":\"...with [underlined: ...]...\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[transitions]\n- Passage with ______ blank connecting two ideas. All 4 choices grammatically correct; distinction is logical.\n- Cover: contrast, addition, cause/effect, example, sequence, concession\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"transitions\",\"passage\":\"...______ ...\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[rhetorical_synthesis]\n- Notes format: \"While researching [topic], a student has taken these notes:\\n- Note 1\\n- Note 2\\n- Note 3\"\n- Specific rhetorical goal stated. All 4 choices are grammatical sentences using the notes; only 1 achieves the goal.\n- Goals: emphasize difference, highlight finding, introduce topic, summarize conclusion, present cause-effect\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"rhetorical_synthesis\",\"passage\":\"...\",\"prompt\":\"The student wants to [goal]. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\nOutput ONLY a valid JSON array of EXACTLY 3 objects. No markdown. Count them before outputting.", "text": "[{\"id\":\"sat99-rw-m2-q31\",\"number\":31,\"type\":\"rhetorical_synthesis\",\"passage\":\"While researching the foraging behavior of bumblebees, a student has taken the following notes:\\n- Bumblebees (genus $Bombus$) visit a wide variety of flowering plants.\\n- A 2019 study by Lars Chittka observed bumblebees in an enclosed arena with artificial flowers.\\n- Some flowers offered high-concentration sucrose solutions; others offered low-concentration solutions.\\n- Bumblebees in the study revisited high-concentration flowers significantly more often than low-concentration ones.\\n- Chittka concluded that bumblebees can distinguish between reward magnitudes and adjust foraging accordingly.\",\"prompt\":\"The student wants to present the main finding of Chittka's 2019 study. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"Bumblebees, members of the genus $Bombus$, are known to visit a wide variety of flowering plants in natural environments.\",\"B\":\"In a 2019 study, Lars Chittka observed bumblebees in an enclosed arena containing artificial flowers with varying sucrose concentrations.\",\"C\":\"In Chittka's 2019 study, bumblebees revisited artificial flowers offering high-concentration sucrose solutions significantly more often than those offering low-concentration solutions.\",\"D\":\"Chittka's research suggests that the foraging behavior of bumblebees is shaped by their ability to perceive features of their environment.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m2-q32\",\"number\":32,\"type\":\"rhetorical_synthesis\",\"passage\":\"While researching the Italian Renaissance painter Sofonisba Anguissola, a student has taken the following notes:\\n- Sofonisba Anguissola (c. 1532–1625) was a Northern Italian painter from Cremona.\\n- She trained under Bernardino Campi, an unusual arrangement for women of her era.\\n- Her 1555 painting $The$ $Chess$ $Game$ depicts her three sisters playing chess.\\n- $The$ $Chess$ $Game$ is considered an early example of a genre scene featuring identifiable individuals.\\n- Most genre scenes of the period depicted anonymous, generalized figures rather than specific people.\",\"prompt\":\"The student wants to emphasize a way in which Anguissola's $The$ $Chess$ $Game$ differed from typical genre scenes of its time. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"Sofonisba Anguissola, who lived from approximately 1532 to 1625, was a painter from the Northern Italian city of Cremona.\",\"B\":\"Whereas most genre scenes of the period depicted anonymous, generalized figures, Anguissola's 1555 painting $The$ $Chess$ $Game$ portrayed her three sisters as identifiable individuals.\",\"C\":\"Anguissola trained under the painter Bernardino Campi, an arrangement that was unusual for women during the Italian Renaissance.\",\"D\":\"Anguissola's $The$ $Chess$ $Game$, completed in 1555, depicts her three sisters engaged in a game of chess.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m2-q33\",\"number\":33,\"type\":\"rhetorical_synthesis\",\"passage\":\"While researching the formation of Saturn's rings, a student has taken the following notes:\\n- Saturn's rings are composed primarily of water ice particles ranging from micrometers to meters in size.\\n- One hypothesis proposes the rings formed roughly 4.5 billion years ago, alongside Saturn itself.\\n- A 2019 analysis of Cassini spacecraft data by Iess and colleagues estimated the rings' total mass.\\n- Based on the low mass and minimal contamination by interplanetary dust, Iess and colleagues concluded the rings are likely only 10 to 100 million years old.\\n- This estimate suggests the rings may have formed long after the planet itself, possibly from the breakup of an icy moon.\",\"prompt\":\"The student wants to introduce the topic of the research to an audience unfamiliar with Saturn's rings. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"Saturn's rings, which are composed primarily of water ice particles varying widely in size, have been the subject of recent scientific investigation into their age and origin.\",\"B\":\"Iess and colleagues analyzed Cassini spacecraft data in 2019 to estimate the total mass of Saturn's rings.\",\"C\":\"The low mass and minimal dust contamination of Saturn's rings led Iess and colleagues to conclude that the rings are 10 to 100 million years old.\",\"D\":\"Saturn's rings may have formed from the breakup of an icy moon long after Saturn itself coalesced roughly 4.5 billion years ago.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"A\"}]" } ], "results": [ { "attempts": 1, "index": 0, "output": { "costUsd": 0.08664, "model": "opus", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 5 SAT Reading & Writing questions.\nType: vocabulary_in_context | Test: 99 | Unit: 1 | Start Q#: 1\n\nMANDATORY: Your JSON array MUST contain exactly 5 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nFollow the instructions for your type:\n\n[vocabulary_in_context]\n- 2-3 sentence passage with ______ blank from specific academic domains (molecular biology, Renaissance art, behavioral economics, astrophysics, political philosophy)\n- 4 college-level word choices from the same semantic field\n- Correct word is most PRECISE for context, not most common synonym\n- ID format: sat{testNumber}-rw-m{mod}-q{N}\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"vocabulary_in_context\",\"passage\":\"...with ______...\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[text_comprehension]\n- 3-8 sentence passage from published-quality source (journal, novel, essay) with specific names/dates/findings\n- Mix subtypes: main idea (2-3), author's purpose (1-2), inference (2-3), tone (0-1), structure (1-2)\n- Topics: natural science, social science, literature, history, arts/humanities (use each)\n- Distractors: surface-level misreading, wrong scope, opposite interpretation\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"text_comprehension\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[data_analysis]\n- Passage describing realistic data/study results requiring INTERPRETATION, not lookup\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Data patterns: diminishing returns, optimal values, conditional relationships\n- Distractors: wrong column, wrong trend direction, correlation vs causation confusion\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"data_analysis\",\"passage\":\"...\",\"visual\":{...},\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[claim_support]\n- Mix: \"Which finding supports/undermines the claim?\" AND \"Which quotation illustrates the claim?\"\n- Claims must be specific and testable. Distractors are topically related but don't address the specific mechanism.\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"claim_support\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[standard_english_conventions]\n- Passage with [underlined: text here] portion. Student picks best replacement.\n- Cover: subject-verb agreement, pronouns, verb tense, parallel structure, modifiers, commas, semicolons, apostrophes, sentence boundaries\n- \"NO CHANGE\" (choice A) correct ~25% of the time, NOT always\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"standard_english_conventions\",\"passage\":\"...with [underlined: ...]...\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[transitions]\n- Passage with ______ blank connecting two ideas. All 4 choices grammatically correct; distinction is logical.\n- Cover: contrast, addition, cause/effect, example, sequence, concession\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"transitions\",\"passage\":\"...______ ...\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[rhetorical_synthesis]\n- Notes format: \"While researching [topic], a student has taken these notes:\\n- Note 1\\n- Note 2\\n- Note 3\"\n- Specific rhetorical goal stated. All 4 choices are grammatical sentences using the notes; only 1 achieves the goal.\n- Goals: emphasize difference, highlight finding, introduce topic, summarize conclusion, present cause-effect\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"rhetorical_synthesis\",\"passage\":\"...\",\"prompt\":\"The student wants to [goal]. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\nOutput ONLY a valid JSON array of EXACTLY 5 objects. No markdown. Count them before outputting.", "text": "[{\"id\":\"sat99-rw-m1-q1\",\"number\":1,\"type\":\"vocabulary_in_context\",\"passage\":\"In his 1503 treatise on proportion, the Florentine sculptor argued that classical statuary achieved its effects not through rigid adherence to mathematical ratios but through subtle departures from them. A slight elongation of the torso, a marginal tilt of the head—these deviations, he claimed, ______ the figure with a vitality that strict symmetry could never produce.\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"imbued\",\"B\":\"contaminated\",\"C\":\"stocked\",\"D\":\"appointed\"},\"answer_format\":\"mcq\",\"correct_answer\":\"A\"},{\"id\":\"sat99-rw-m1-q2\",\"number\":2,\"type\":\"vocabulary_in_context\",\"passage\":\"Behavioral economist Sendhil Mullainathan has argued that scarcity does not merely shape decisions—it ______ cognition itself. When the poor must constantly calculate trade-offs that the wealthy never notice, mental bandwidth is consumed by the arithmetic of survival, leaving fewer resources for long-term planning.\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"designates\",\"B\":\"restructures\",\"C\":\"transcribes\",\"D\":\"deputizes\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m1-q3\",\"number\":3,\"type\":\"vocabulary_in_context\",\"passage\":\"The discovery of FOXP2, often misleadingly dubbed the ''language gene,'' has been widely ______ in popular science writing. Though mutations in the gene do disrupt speech production in humans, researchers caution that it is one component of a vast regulatory network rather than a singular switch governing linguistic capacity.\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"repudiated\",\"B\":\"curtailed\",\"C\":\"oversimplified\",\"D\":\"fabricated\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m1-q4\",\"number\":4,\"type\":\"vocabulary_in_context\",\"passage\":\"Because Population III stars—the hypothetical first generation formed from primordial hydrogen and helium—have never been directly observed, their properties must be ______ from indirect evidence such as the chemical signatures of ancient, metal-poor stars still visible in the Milky Way's halo.\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"extracted\",\"B\":\"inferred\",\"C\":\"fabricated\",\"D\":\"disclosed\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m1-q5\",\"number\":5,\"type\":\"vocabulary_in_context\",\"passage\":\"In $\\\\textit{On Liberty}$, John Stuart Mill worried that majority opinion, when unchecked, could exert a form of coercion more insidious than any statute. Public disapproval, he argued, could ______ dissent so thoroughly that heterodox ideas would wither before they could even be articulated.\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"stifle\",\"B\":\"rebuke\",\"C\":\"dispute\",\"D\":\"evade\"},\"answer_format\":\"mcq\",\"correct_answer\":\"A\"}]" }, "status": "ok" }, { "attempts": 1, "index": 1, "output": { "costUsd": 0.280965, "model": "opus", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 8 SAT Reading & Writing questions.\nType: text_comprehension | Test: 99 | Unit: 1 | Start Q#: 6\n\nMANDATORY: Your JSON array MUST contain exactly 8 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nFollow the instructions for your type:\n\n[vocabulary_in_context]\n- 2-3 sentence passage with ______ blank from specific academic domains (molecular biology, Renaissance art, behavioral economics, astrophysics, political philosophy)\n- 4 college-level word choices from the same semantic field\n- Correct word is most PRECISE for context, not most common synonym\n- ID format: sat{testNumber}-rw-m{mod}-q{N}\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"vocabulary_in_context\",\"passage\":\"...with ______...\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[text_comprehension]\n- 3-8 sentence passage from published-quality source (journal, novel, essay) with specific names/dates/findings\n- Mix subtypes: main idea (2-3), author's purpose (1-2), inference (2-3), tone (0-1), structure (1-2)\n- Topics: natural science, social science, literature, history, arts/humanities (use each)\n- Distractors: surface-level misreading, wrong scope, opposite interpretation\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"text_comprehension\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[data_analysis]\n- Passage describing realistic data/study results requiring INTERPRETATION, not lookup\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Data patterns: diminishing returns, optimal values, conditional relationships\n- Distractors: wrong column, wrong trend direction, correlation vs causation confusion\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"data_analysis\",\"passage\":\"...\",\"visual\":{...},\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[claim_support]\n- Mix: \"Which finding supports/undermines the claim?\" AND \"Which quotation illustrates the claim?\"\n- Claims must be specific and testable. Distractors are topically related but don't address the specific mechanism.\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"claim_support\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[standard_english_conventions]\n- Passage with [underlined: text here] portion. Student picks best replacement.\n- Cover: subject-verb agreement, pronouns, verb tense, parallel structure, modifiers, commas, semicolons, apostrophes, sentence boundaries\n- \"NO CHANGE\" (choice A) correct ~25% of the time, NOT always\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"standard_english_conventions\",\"passage\":\"...with [underlined: ...]...\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[transitions]\n- Passage with ______ blank connecting two ideas. All 4 choices grammatically correct; distinction is logical.\n- Cover: contrast, addition, cause/effect, example, sequence, concession\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"transitions\",\"passage\":\"...______ ...\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[rhetorical_synthesis]\n- Notes format: \"While researching [topic], a student has taken these notes:\\n- Note 1\\n- Note 2\\n- Note 3\"\n- Specific rhetorical goal stated. All 4 choices are grammatical sentences using the notes; only 1 achieves the goal.\n- Goals: emphasize difference, highlight finding, introduce topic, summarize conclusion, present cause-effect\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"rhetorical_synthesis\",\"passage\":\"...\",\"prompt\":\"The student wants to [goal]. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\nOutput ONLY a valid JSON array of EXACTLY 8 objects. No markdown. Count them before outputting.", "text": "[{\"id\":\"sat99-rw-m1-q6\",\"number\":6,\"type\":\"text_comprehension\",\"passage\":\"In her 2019 study of cooperative breeding in pied babblers (Turdoides bicolor), zoologist Amanda Ridley documented that subordinate adults who had previously received vocal encouragement from dominant pair members were significantly more likely to provision unrelated nestlings than those who had not. Ridley argues that this pattern complicates the standard kin-selection account, which holds that helping behaviors evolve primarily because they benefit close genetic relatives. The vocal exchanges, she suggests, function less as instructions than as a kind of social currency that helpers accumulate and later redeem through opportunities to feed and be associated with the brood.\",\"prompt\":\"Which choice best states the main idea of the text?\",\"choices\":{\"A\":\"Ridley's research demonstrates that kin selection plays no measurable role in the cooperative behavior of pied babblers.\",\"B\":\"Ridley's findings suggest that social interactions, rather than genetic relatedness alone, may shape helping behavior in pied babblers.\",\"C\":\"Ridley discovered that dominant pied babblers use vocal commands to direct unrelated subordinates toward specific nestlings.\",\"D\":\"Ridley concludes that subordinate pied babblers provision nestlings primarily to gain reproductive opportunities of their own.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m1-q7\",\"number\":7,\"type\":\"text_comprehension\",\"passage\":\"The following text is adapted from Edith Wharton's 1905 novel The House of Mirth. Lily Bart, a young woman of limited means, has just encountered an acquaintance at a train station.\\n\\nSelden paused in surprise. In the afternoon rush of the Grand Central Station his eyes had been refreshed by the sight of Miss Lily Bart. It was a Monday in early September, and he was returning to his work from a hurried dip into the country; but what was Miss Bart doing in town at that season? If she had appeared to be catching a train, he might have inferred that he had come on her in the act of transition between one and another of the country-houses which disputed her presence after the close of the Newport season; but her desultory air perplexed him. She stood apart from the crowd, letting it drift by her to the platform or the street, and wearing an air of irresolution which might, as he surmised, be the mask of a very definite purpose.\",\"prompt\":\"Which choice best describes the function of the final sentence in the text as a whole?\",\"choices\":{\"A\":\"It resolves Selden's earlier confusion by clarifying Lily's reasons for being at the station.\",\"B\":\"It contrasts Lily's outward behavior with the more chaotic behavior of the surrounding crowd.\",\"C\":\"It introduces Selden's suspicion that Lily's apparent indecision may conceal a deliberate intention.\",\"D\":\"It suggests that Lily is unaware of Selden's presence and is preoccupied with private worries.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m1-q8\",\"number\":8,\"type\":\"text_comprehension\",\"passage\":\"Art historian Patricia Fortini Brown has long maintained that fifteenth-century Venetian painters approached classical antiquity differently from their Florentine contemporaries. Where Florentine artists such as Mantegna treated ancient ruins and inscriptions as authoritative models to be reconstructed with archaeological precision, Venetian painters like Gentile Bellini incorporated antique fragments into compositions that subordinated historical accuracy to civic narrative. For Brown, the resulting images are not failures of scholarship but deliberate acts of appropriation: Venice, lacking a Roman past of its own, fashioned one by absorbing the relics of others into a visual rhetoric of imperial continuity.\",\"prompt\":\"As used in the text, the phrase \\\"deliberate acts of appropriation\\\" most nearly suggests that Venetian painters\",\"choices\":{\"A\":\"borrowed antique elements carelessly because they lacked access to reliable archaeological sources.\",\"B\":\"intentionally repurposed classical material to construct a flattering historical identity for their city.\",\"C\":\"competed with Florentine artists by producing more historically accurate reconstructions of Roman monuments.\",\"D\":\"rejected classical sources entirely in favor of subjects drawn from Venetian civic life.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m1-q9\",\"number\":9,\"type\":\"text_comprehension\",\"passage\":\"Recent measurements from the Parker Solar Probe have revealed that the solar wind is punctuated by sharp, S-shaped reversals in the magnetic field, which mission scientists have termed \\\"switchbacks.\\\" Although these structures had been glimpsed by earlier spacecraft, their abundance close to the Sun was not anticipated by prevailing models of coronal heating. Justin Kasper, the mission's principal investigator, cautions that it remains unclear whether switchbacks are generated near the solar surface by small-scale magnetic reconnection events or arise farther out through turbulent interactions in the expanding wind. Distinguishing between these possibilities, he notes, will require coordinated observations from probes operating at different distances from the Sun.\",\"prompt\":\"Based on the text, what can most reasonably be inferred about the origin of switchbacks?\",\"choices\":{\"A\":\"Researchers have determined that switchbacks form primarily through magnetic reconnection at the Sun's surface.\",\"B\":\"Researchers consider switchbacks too rare near the Sun to be explained by current coronal heating models.\",\"C\":\"Researchers do not yet have observations sufficient to decide between competing explanations of how switchbacks form.\",\"D\":\"Researchers believe switchbacks are an artifact of the Parker Solar Probe's instruments rather than a genuine solar phenomenon.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m1-q10\",\"number\":10,\"type\":\"text_comprehension\",\"passage\":\"In The Federalist No. 10, James Madison concedes that factions, which he defines as groups united by interests adverse to the rights of others or to the permanent interests of the community, are an unavoidable feature of free societies. To suppress factions, he argues, would require either destroying the liberty that nourishes them or imposing a uniformity of opinion that is impossible among reasoning beings. Madison therefore turns from the question of preventing factions to that of controlling their effects, contending that an extended republic, by multiplying the variety of interests it contains, makes the formation of an oppressive majority less probable than it would be in a small, homogeneous democracy.\",\"prompt\":\"Which choice best describes the overall structure of the text?\",\"choices\":{\"A\":\"It defines a political problem, dismisses two proposed solutions, and advances an alternative remedy.\",\"B\":\"It presents a historical example, identifies its weaknesses, and proposes a return to earlier practices.\",\"C\":\"It introduces a controversial claim, surveys objections to it, and concludes that the claim cannot be defended.\",\"D\":\"It compares two competing forms of government and recommends a compromise between them.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"A\"},{\"id\":\"sat99-rw-m1-q11\",\"number\":11,\"type\":\"text_comprehension\",\"passage\":\"The following text is adapted from a 1903 essay by W. E. B. Du Bois.\\n\\nBetween me and the other world there is ever an unasked question: unasked by some through feelings of delicacy; by others through the difficulty of rightly framing it. All, nevertheless, flutter round it. They approach me in a half-hesitant sort of way, eye me curiously or compassionately, and then, instead of saying directly, How does it feel to be a problem? they say, I know an excellent colored man in my town; or, I fought at Mechanicsville; or, Do not these Southern outrages make your blood boil? At these I smile, or am interested, or reduce the boiling to a simmer, as the occasion may require. To the real question, How does it feel to be a problem? I answer seldom a word.\",\"prompt\":\"Which choice best describes the speaker's tone in the text?\",\"choices\":{\"A\":\"Wry restraint mingled with quiet weariness\",\"B\":\"Defiant outrage barely contained by politeness\",\"C\":\"Cheerful detachment about a trivial misunderstanding\",\"D\":\"Hopeful conviction that misunderstandings will soon be resolved\"},\"answer_format\":\"mcq\",\"correct_answer\":\"A\"},{\"id\":\"sat99-rw-m1-q12\",\"number\":12,\"type\":\"text_comprehension\",\"passage\":\"Behavioral economist Sendhil Mullainathan and his collaborators have argued that the experience of scarcity, whether of money, time, or calories, alters cognition in ways that conventional economic models fail to capture. In a series of experiments, participants asked to manage tight budgets performed better on tasks directly related to those budgets but worse on unrelated tests of fluid intelligence and self-control. Mullainathan describes this trade-off as a kind of cognitive \\\"tax\\\" levied by scarcity itself, one that may help explain why interventions designed to alleviate poverty often falter when they demand additional planning, paperwork, or attention from the very people they aim to assist.\",\"prompt\":\"Why does the author most likely describe scarcity as a cognitive \\\"tax\\\"?\",\"choices\":{\"A\":\"To suggest that governments should compensate individuals financially for the mental costs of being poor\",\"B\":\"To emphasize that scarcity imposes mental costs that can undermine programs requiring sustained effort from beneficiaries\",\"C\":\"To argue that conventional economic models accurately predict how people respond to financial constraints\",\"D\":\"To imply that people experiencing scarcity perform worse on every cognitive task than wealthier people do\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m1-q13\",\"number\":13,\"type\":\"text_comprehension\",\"passage\":\"Marine biologist Roger Payne's 1971 recordings of humpback whale songs are often credited with galvanizing public support for the international moratorium on commercial whaling adopted a decade later. Yet the recordings' influence cannot be attributed to scientific argument alone. By releasing the songs as a commercial album rather than confining them to academic journals, Payne placed the voices of whales within the same cultural space as popular music, inviting listeners to perceive the animals as expressive beings rather than as units of harvestable biomass. The conservation movement that followed drew as much from this aesthetic reframing as from the demographic data that researchers were simultaneously compiling.\",\"prompt\":\"Which choice best states the main purpose of the text?\",\"choices\":{\"A\":\"To document the scientific methods Payne used to record and analyze humpback whale vocalizations\",\"B\":\"To argue that the moratorium on commercial whaling was adopted primarily because of advances in population data\",\"C\":\"To suggest that Payne's recordings shaped public attitudes toward whales in part by reframing them aesthetically rather than only scientifically\",\"D\":\"To criticize the conservation movement for relying on emotional appeals rather than on rigorous scientific evidence\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"}]" }, "status": "ok" }, { "attempts": 1, "index": 2, "output": { "costUsd": 0.15984, "model": "opus", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 4 SAT Reading & Writing questions.\nType: data_analysis | Test: 99 | Unit: 1 | Start Q#: 14\n\nMANDATORY: Your JSON array MUST contain exactly 4 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nFollow the instructions for your type:\n\n[vocabulary_in_context]\n- 2-3 sentence passage with ______ blank from specific academic domains (molecular biology, Renaissance art, behavioral economics, astrophysics, political philosophy)\n- 4 college-level word choices from the same semantic field\n- Correct word is most PRECISE for context, not most common synonym\n- ID format: sat{testNumber}-rw-m{mod}-q{N}\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"vocabulary_in_context\",\"passage\":\"...with ______...\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[text_comprehension]\n- 3-8 sentence passage from published-quality source (journal, novel, essay) with specific names/dates/findings\n- Mix subtypes: main idea (2-3), author's purpose (1-2), inference (2-3), tone (0-1), structure (1-2)\n- Topics: natural science, social science, literature, history, arts/humanities (use each)\n- Distractors: surface-level misreading, wrong scope, opposite interpretation\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"text_comprehension\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[data_analysis]\n- Passage describing realistic data/study results requiring INTERPRETATION, not lookup\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Data patterns: diminishing returns, optimal values, conditional relationships\n- Distractors: wrong column, wrong trend direction, correlation vs causation confusion\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"data_analysis\",\"passage\":\"...\",\"visual\":{...},\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[claim_support]\n- Mix: \"Which finding supports/undermines the claim?\" AND \"Which quotation illustrates the claim?\"\n- Claims must be specific and testable. Distractors are topically related but don't address the specific mechanism.\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"claim_support\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[standard_english_conventions]\n- Passage with [underlined: text here] portion. Student picks best replacement.\n- Cover: subject-verb agreement, pronouns, verb tense, parallel structure, modifiers, commas, semicolons, apostrophes, sentence boundaries\n- \"NO CHANGE\" (choice A) correct ~25% of the time, NOT always\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"standard_english_conventions\",\"passage\":\"...with [underlined: ...]...\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[transitions]\n- Passage with ______ blank connecting two ideas. All 4 choices grammatically correct; distinction is logical.\n- Cover: contrast, addition, cause/effect, example, sequence, concession\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"transitions\",\"passage\":\"...______ ...\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[rhetorical_synthesis]\n- Notes format: \"While researching [topic], a student has taken these notes:\\n- Note 1\\n- Note 2\\n- Note 3\"\n- Specific rhetorical goal stated. All 4 choices are grammatical sentences using the notes; only 1 achieves the goal.\n- Goals: emphasize difference, highlight finding, introduce topic, summarize conclusion, present cause-effect\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"rhetorical_synthesis\",\"passage\":\"...\",\"prompt\":\"The student wants to [goal]. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\nOutput ONLY a valid JSON array of EXACTLY 4 objects. No markdown. Count them before outputting.", "text": "[{\"id\":\"sat99-rw-m1-q14\",\"number\":14,\"type\":\"data_analysis\",\"passage\":\"Soil ecologist Maren Holst investigated how earthworm density affects the decomposition rate of leaf litter in temperate deciduous forests. She established experimental plots with varying numbers of earthworms per square meter and measured the percentage of leaf mass lost over a 90-day period. Holst hypothesized that decomposition would increase with earthworm density but that the rate of increase would diminish at higher densities because of resource competition among the worms.\",\"visual\":{\"type\":\"table\",\"title\":\"Leaf Litter Decomposition by Earthworm Density (90 Days)\",\"headers\":[\"Earthworms per m²\",\"Leaf Mass Lost (%)\",\"Increase from Previous Level (%)\"],\"rows\":[[\"0\",\"12\",\"—\"],[\"20\",\"28\",\"16\"],[\"40\",\"41\",\"13\"],[\"60\",\"49\",\"8\"],[\"80\",\"53\",\"4\"]]},\"prompt\":\"Which choice best describes data from the table that support Holst's hypothesis?\",\"choices\":{\"A\":\"Decomposition was highest at 80 earthworms per m², and the increase from the previous density level was the smallest observed.\",\"B\":\"Decomposition increased steadily from 0 to 80 earthworms per m², and the increase from each previous level was identical across densities.\",\"C\":\"Decomposition was lowest at 0 earthworms per m², and the increase from 0 to 20 earthworms per m² was smaller than the increase from 60 to 80.\",\"D\":\"Decomposition peaked at 40 earthworms per m², after which leaf mass lost declined as density increased further.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"A\"},{\"id\":\"sat99-rw-m1-q15\",\"number\":15,\"type\":\"data_analysis\",\"passage\":\"A team of behavioral economists led by Rajiv Chandra examined whether the timing of a charitable appeal influences donation amounts. Participants were randomly assigned to receive a solicitation email either on the morning of payday, three days after payday, or ten days after payday. The researchers recorded the mean donation in each condition as well as the percentage of recipients who donated at all.\",\"visual\":{\"type\":\"table\",\"title\":\"Donation Behavior by Email Timing\",\"headers\":[\"Timing of Appeal\",\"Donation Rate (%)\",\"Mean Donation ($)\"],\"rows\":[[\"Morning of payday\",\"34\",\"22.50\"],[\"3 days after payday\",\"29\",\"18.75\"],[\"10 days after payday\",\"18\",\"11.20\"]]},\"prompt\":\"Which choice most accurately describes the relationship between appeal timing and donor behavior shown in the table?\",\"choices\":{\"A\":\"As the delay after payday increased, the donation rate rose while the mean donation amount fell.\",\"B\":\"As the delay after payday increased, both the donation rate and the mean donation amount declined.\",\"C\":\"The donation rate was highest ten days after payday, but the mean donation was highest on the morning of payday.\",\"D\":\"Appeals sent three days after payday produced both the highest donation rate and the highest mean donation.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m1-q16\",\"number\":16,\"type\":\"data_analysis\",\"passage\":\"Materials scientist Inés Beltrán tested the tensile strength of a novel bio-based polymer at several curing temperatures to determine an optimal processing condition. She noted that insufficient curing leaves polymer chains weakly bonded, whereas excessive heat can degrade the chains and reduce strength. Beltrán predicted that tensile strength would rise with temperature up to a threshold and then fall beyond it.\",\"visual\":{\"type\":\"table\",\"title\":\"Tensile Strength of Bio-Based Polymer by Curing Temperature\",\"headers\":[\"Curing Temperature (°C)\",\"Tensile Strength (MPa)\"],\"rows\":[[\"100\",\"42\"],[\"120\",\"58\"],[\"140\",\"71\"],[\"160\",\"65\"],[\"180\",\"49\"]]},\"prompt\":\"Which choice best describes data from the table that support Beltrán's prediction?\",\"choices\":{\"A\":\"Tensile strength rose continuously from 100 °C to 180 °C, confirming that higher curing temperatures always improve the polymer.\",\"B\":\"Tensile strength was lowest at 100 °C and highest at 180 °C, indicating that thermal degradation did not occur within the tested range.\",\"C\":\"Tensile strength increased from 100 °C to 140 °C and then decreased at 160 °C and 180 °C, indicating a peak within the tested range.\",\"D\":\"Tensile strength remained constant between 120 °C and 160 °C, suggesting that curing temperature has little effect in that interval.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m1-q17\",\"number\":17,\"type\":\"data_analysis\",\"passage\":\"In a study of cognitive load during multitasking, psychologist Theo Abedi asked participants to complete a primary reading comprehension task while performing a secondary task of varying difficulty. He measured reading accuracy (percentage of comprehension questions answered correctly) and reaction time on the secondary task. Abedi wanted to determine whether a moderately demanding secondary task would disrupt primary-task performance less than either a very easy or a very difficult one.\",\"visual\":{\"type\":\"table\",\"title\":\"Performance by Secondary-Task Difficulty\",\"headers\":[\"Secondary-Task Difficulty\",\"Reading Accuracy (%)\",\"Secondary-Task Reaction Time (ms)\"],\"rows\":[[\"Very easy\",\"74\",\"310\"],[\"Moderate\",\"88\",\"425\"],[\"Difficult\",\"61\",\"690\"]]},\"prompt\":\"Which choice best describes data from the table that are relevant to Abedi's research question?\",\"choices\":{\"A\":\"Reading accuracy was lower in the difficult condition than in the very easy condition, showing that harder secondary tasks always impair primary-task performance.\",\"B\":\"Reading accuracy was higher in the moderate condition than in either the very easy or the difficult condition, consistent with the idea that a moderately demanding secondary task disrupts reading least.\",\"C\":\"Reaction time on the secondary task was shortest in the very easy condition, indicating that participants focused most effectively on reading when the secondary task was simplest.\",\"D\":\"Reading accuracy and secondary-task reaction time both increased as difficulty rose, suggesting that the two measures track each other across conditions.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"}]" }, "status": "ok" }, { "attempts": 1, "index": 3, "output": { "costUsd": 0.150315, "model": "opus", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 4 SAT Reading & Writing questions.\nType: claim_support | Test: 99 | Unit: 1 | Start Q#: 18\n\nMANDATORY: Your JSON array MUST contain exactly 4 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nFollow the instructions for your type:\n\n[vocabulary_in_context]\n- 2-3 sentence passage with ______ blank from specific academic domains (molecular biology, Renaissance art, behavioral economics, astrophysics, political philosophy)\n- 4 college-level word choices from the same semantic field\n- Correct word is most PRECISE for context, not most common synonym\n- ID format: sat{testNumber}-rw-m{mod}-q{N}\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"vocabulary_in_context\",\"passage\":\"...with ______...\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[text_comprehension]\n- 3-8 sentence passage from published-quality source (journal, novel, essay) with specific names/dates/findings\n- Mix subtypes: main idea (2-3), author's purpose (1-2), inference (2-3), tone (0-1), structure (1-2)\n- Topics: natural science, social science, literature, history, arts/humanities (use each)\n- Distractors: surface-level misreading, wrong scope, opposite interpretation\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"text_comprehension\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[data_analysis]\n- Passage describing realistic data/study results requiring INTERPRETATION, not lookup\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Data patterns: diminishing returns, optimal values, conditional relationships\n- Distractors: wrong column, wrong trend direction, correlation vs causation confusion\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"data_analysis\",\"passage\":\"...\",\"visual\":{...},\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[claim_support]\n- Mix: \"Which finding supports/undermines the claim?\" AND \"Which quotation illustrates the claim?\"\n- Claims must be specific and testable. Distractors are topically related but don't address the specific mechanism.\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"claim_support\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[standard_english_conventions]\n- Passage with [underlined: text here] portion. Student picks best replacement.\n- Cover: subject-verb agreement, pronouns, verb tense, parallel structure, modifiers, commas, semicolons, apostrophes, sentence boundaries\n- \"NO CHANGE\" (choice A) correct ~25% of the time, NOT always\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"standard_english_conventions\",\"passage\":\"...with [underlined: ...]...\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[transitions]\n- Passage with ______ blank connecting two ideas. All 4 choices grammatically correct; distinction is logical.\n- Cover: contrast, addition, cause/effect, example, sequence, concession\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"transitions\",\"passage\":\"...______ ...\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[rhetorical_synthesis]\n- Notes format: \"While researching [topic], a student has taken these notes:\\n- Note 1\\n- Note 2\\n- Note 3\"\n- Specific rhetorical goal stated. All 4 choices are grammatical sentences using the notes; only 1 achieves the goal.\n- Goals: emphasize difference, highlight finding, introduce topic, summarize conclusion, present cause-effect\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"rhetorical_synthesis\",\"passage\":\"...\",\"prompt\":\"The student wants to [goal]. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\nOutput ONLY a valid JSON array of EXACTLY 4 objects. No markdown. Count them before outputting.", "text": "[{\"id\":\"sat99-rw-m1-q18\",\"number\":18,\"type\":\"claim_support\",\"passage\":\"A researcher studying the cognitive development of corvids (a family of birds including crows and ravens) has proposed that New Caledonian crows possess a form of causal reasoning that extends beyond simple trial-and-error learning. Specifically, the researcher claims that these birds can mentally simulate the physical consequences of an action before executing it, allowing them to solve novel tool-use problems on the first attempt without prior exposure.\",\"prompt\":\"Which finding, if true, would most directly support the researcher's claim?\",\"choices\":{\"A\":\"In laboratory settings, New Caledonian crows that observed other crows retrieving food with hooked sticks were subsequently able to use similar tools themselves after several practice sessions.\",\"B\":\"When presented with a novel apparatus requiring them to drop stones into a water-filled tube to raise a floating reward, naive New Caledonian crows selected stones of appropriate size and density on their first attempt, ignoring objects that would float.\",\"C\":\"New Caledonian crows in the wild have been documented manufacturing hooked tools from pandanus leaves, a behavior that varies regionally and appears to be transmitted across generations.\",\"D\":\"Neuroanatomical studies reveal that New Caledonian crows possess a higher density of neurons in the pallium than other corvid species of comparable body size.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m1-q19\",\"number\":19,\"type\":\"claim_support\",\"passage\":\"In her 2021 study of urban microclimates, geographer Lena Abernathy argues that the cooling effect of small neighborhood parks on surrounding city blocks is disproportionately large relative to their area—a phenomenon she attributes not to the parks' vegetation alone but to the way they disrupt the continuous surface of heat-absorbing pavement and buildings that characterize dense urban cores.\",\"prompt\":\"Which finding from Abernathy's study would most directly support her specific explanation for the cooling effect?\",\"choices\":{\"A\":\"Neighborhood parks measuring less than one hectare were associated with average temperature reductions of 1.8°C in adjacent blocks during summer afternoons.\",\"B\":\"Tree-canopy coverage within the parks correlated positively with the magnitude of temperature reduction observed in surrounding areas.\",\"C\":\"Paved plazas of equivalent size to the parks, when inserted into dense urban blocks, produced measurable cooling effects roughly 60 percent as large as those of vegetated parks.\",\"D\":\"Residents living within 200 meters of a neighborhood park reported lower rates of heat-related illness than residents living farther away during extreme heat events.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m1-q20\",\"number\":20,\"type\":\"claim_support\",\"passage\":\"In his critical study of the Victorian novel, literary scholar Marcus Holbrook contends that Elizabeth Gaskell's \\\"North and South\\\" (1855) departs from the conventions of the industrial novel by granting its female protagonist, Margaret Hale, a role as economic arbiter rather than as a passive moral influence. Holbrook argues that Margaret's decisions shape the material fortunes of the novel's central mill, a reversal of the typical Victorian pattern in which women inspire male action without directly determining business outcomes.\",\"prompt\":\"Which quotation from \\\"North and South\\\" would most effectively illustrate Holbrook's claim?\",\"choices\":{\"A\":\"\\\"Margaret felt, as she watched the crowds streaming from the mill gates, a sorrow she could scarcely name, a pity for the weary faces that seemed to reproach her own ease.\\\"\",\"B\":\"\\\"Mr. Thornton's manner toward her had grown gentler since their last conversation, and she was aware that her words had, in some measure, softened his views on the conduct of his workmen.\\\"\",\"C\":\"\\\"It was Margaret's capital, inherited from Mr. Bell and offered upon terms of her own devising, that permitted Thornton to reopen Marlborough Mills when no bank in Milton would advance him a shilling.\\\"\",\"D\":\"\\\"She thought often of the North, of its smoke and clangor, and wondered whether the gentleness of the South had not unfitted her for a world in which men like Thornton labored and contended.\\\"\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m1-q21\",\"number\":21,\"type\":\"claim_support\",\"passage\":\"Archaeologist Priya Chandran has argued that the residents of Çatalhöyük, a Neolithic settlement in southern Anatolia occupied from roughly 7100 to 5950 BCE, maintained a remarkably egalitarian social structure throughout the settlement's occupation, with no single household accumulating significantly greater wealth or status than its neighbors. Critics have suggested that this apparent egalitarianism may reflect the limitations of the archaeological record rather than the actual social organization of the community.\",\"prompt\":\"Which finding, if true, would most directly undermine Chandran's claim?\",\"choices\":{\"A\":\"Isotopic analysis of human remains from Çatalhöyük indicates that dietary composition varied little across individuals, suggesting broadly similar access to food resources.\",\"B\":\"Excavations at a contemporaneous settlement 200 kilometers to the east have revealed clear evidence of status differentiation, including elite burials with imported obsidian goods.\",\"C\":\"Detailed analysis of the distribution of rare imported materials across Çatalhöyük households reveals that a small cluster of dwellings consistently contained ten times the quantity found elsewhere in the settlement.\",\"D\":\"The layout of Çatalhöyük, with its densely packed houses sharing walls and accessed through roof openings, has been interpreted as evidence of strong communal identity.\",\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"}]" }, "status": "ok" }, { "attempts": 1, "index": 4, "output": { "costUsd": 0.08124, "model": "opus", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 5 SAT Reading & Writing questions.\nType: standard_english_conventions | Test: 99 | Unit: 1 | Start Q#: 22\n\nMANDATORY: Your JSON array MUST contain exactly 5 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nFollow the instructions for your type:\n\n[vocabulary_in_context]\n- 2-3 sentence passage with ______ blank from specific academic domains (molecular biology, Renaissance art, behavioral economics, astrophysics, political philosophy)\n- 4 college-level word choices from the same semantic field\n- Correct word is most PRECISE for context, not most common synonym\n- ID format: sat{testNumber}-rw-m{mod}-q{N}\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"vocabulary_in_context\",\"passage\":\"...with ______...\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[text_comprehension]\n- 3-8 sentence passage from published-quality source (journal, novel, essay) with specific names/dates/findings\n- Mix subtypes: main idea (2-3), author's purpose (1-2), inference (2-3), tone (0-1), structure (1-2)\n- Topics: natural science, social science, literature, history, arts/humanities (use each)\n- Distractors: surface-level misreading, wrong scope, opposite interpretation\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"text_comprehension\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[data_analysis]\n- Passage describing realistic data/study results requiring INTERPRETATION, not lookup\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Data patterns: diminishing returns, optimal values, conditional relationships\n- Distractors: wrong column, wrong trend direction, correlation vs causation confusion\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"data_analysis\",\"passage\":\"...\",\"visual\":{...},\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[claim_support]\n- Mix: \"Which finding supports/undermines the claim?\" AND \"Which quotation illustrates the claim?\"\n- Claims must be specific and testable. Distractors are topically related but don't address the specific mechanism.\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"claim_support\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[standard_english_conventions]\n- Passage with [underlined: text here] portion. Student picks best replacement.\n- Cover: subject-verb agreement, pronouns, verb tense, parallel structure, modifiers, commas, semicolons, apostrophes, sentence boundaries\n- \"NO CHANGE\" (choice A) correct ~25% of the time, NOT always\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"standard_english_conventions\",\"passage\":\"...with [underlined: ...]...\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[transitions]\n- Passage with ______ blank connecting two ideas. All 4 choices grammatically correct; distinction is logical.\n- Cover: contrast, addition, cause/effect, example, sequence, concession\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"transitions\",\"passage\":\"...______ ...\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[rhetorical_synthesis]\n- Notes format: \"While researching [topic], a student has taken these notes:\\n- Note 1\\n- Note 2\\n- Note 3\"\n- Specific rhetorical goal stated. All 4 choices are grammatical sentences using the notes; only 1 achieves the goal.\n- Goals: emphasize difference, highlight finding, introduce topic, summarize conclusion, present cause-effect\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"rhetorical_synthesis\",\"passage\":\"...\",\"prompt\":\"The student wants to [goal]. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\nOutput ONLY a valid JSON array of EXACTLY 5 objects. No markdown. Count them before outputting.", "text": "[{\"id\":\"sat99-rw-m1-q22\",\"number\":22,\"type\":\"standard_english_conventions\",\"passage\":\"The sculptor Louise Nevelson, who emigrated from what is now Ukraine to the United States in 1905, became renowned for her monochromatic wooden assemblages; these works, often painted entirely black or white, [underlined: transforms] discarded objects into monumental compositions.\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"transform\",\"C\":\"is transforming\",\"D\":\"transforming\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m1-q23\",\"number\":23,\"type\":\"standard_english_conventions\",\"passage\":\"Marine biologist Sylvia Earle has spent decades advocating for ocean conservation. Her research expeditions have taken her to depths few humans have reached, and [underlined: they're findings have] reshaped public understanding of deep-sea ecosystems.\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"their findings have\",\"C\":\"there findings have\",\"D\":\"they're findings has\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m1-q24\",\"number\":24,\"type\":\"standard_english_conventions\",\"passage\":\"In traditional Japanese $kintsugi$, artisans repair broken pottery with lacquer mixed with powdered gold, silver, or [underlined: platinum, the technique treats] breakage and repair as part of an object's history rather than something to disguise.\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"platinum the technique treats\",\"C\":\"platinum; the technique treats\",\"D\":\"platinum, treating\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m1-q25\",\"number\":25,\"type\":\"standard_english_conventions\",\"passage\":\"Economist Esther Duflo, who shared the 2019 Nobel Prize for her experimental approach to alleviating global poverty, argues that small-scale randomized trials can reveal which interventions actually [underlined: works] in specific local contexts.\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"is working\",\"C\":\"has worked\",\"D\":\"work\"},\"answer_format\":\"mcq\",\"correct_answer\":\"D\"},{\"id\":\"sat99-rw-m1-q26\",\"number\":26,\"type\":\"standard_english_conventions\",\"passage\":\"By the time astronomers confirmed the existence of the first exoplanet orbiting a sun-like star in 1995, decades of speculation about worlds beyond our solar [underlined: system, had] given way to a concrete, observable field of research.\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"system had\",\"C\":\"system; had\",\"D\":\"system: had\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"}]" }, "status": "ok" }, { "attempts": 1, "index": 5, "output": { "costUsd": 0.07734, "model": "opus", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 4 SAT Reading & Writing questions.\nType: transitions | Test: 99 | Unit: 1 | Start Q#: 27\n\nMANDATORY: Your JSON array MUST contain exactly 4 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nFollow the instructions for your type:\n\n[vocabulary_in_context]\n- 2-3 sentence passage with ______ blank from specific academic domains (molecular biology, Renaissance art, behavioral economics, astrophysics, political philosophy)\n- 4 college-level word choices from the same semantic field\n- Correct word is most PRECISE for context, not most common synonym\n- ID format: sat{testNumber}-rw-m{mod}-q{N}\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"vocabulary_in_context\",\"passage\":\"...with ______...\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[text_comprehension]\n- 3-8 sentence passage from published-quality source (journal, novel, essay) with specific names/dates/findings\n- Mix subtypes: main idea (2-3), author's purpose (1-2), inference (2-3), tone (0-1), structure (1-2)\n- Topics: natural science, social science, literature, history, arts/humanities (use each)\n- Distractors: surface-level misreading, wrong scope, opposite interpretation\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"text_comprehension\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[data_analysis]\n- Passage describing realistic data/study results requiring INTERPRETATION, not lookup\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Data patterns: diminishing returns, optimal values, conditional relationships\n- Distractors: wrong column, wrong trend direction, correlation vs causation confusion\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"data_analysis\",\"passage\":\"...\",\"visual\":{...},\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[claim_support]\n- Mix: \"Which finding supports/undermines the claim?\" AND \"Which quotation illustrates the claim?\"\n- Claims must be specific and testable. Distractors are topically related but don't address the specific mechanism.\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"claim_support\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[standard_english_conventions]\n- Passage with [underlined: text here] portion. Student picks best replacement.\n- Cover: subject-verb agreement, pronouns, verb tense, parallel structure, modifiers, commas, semicolons, apostrophes, sentence boundaries\n- \"NO CHANGE\" (choice A) correct ~25% of the time, NOT always\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"standard_english_conventions\",\"passage\":\"...with [underlined: ...]...\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[transitions]\n- Passage with ______ blank connecting two ideas. All 4 choices grammatically correct; distinction is logical.\n- Cover: contrast, addition, cause/effect, example, sequence, concession\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"transitions\",\"passage\":\"...______ ...\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[rhetorical_synthesis]\n- Notes format: \"While researching [topic], a student has taken these notes:\\n- Note 1\\n- Note 2\\n- Note 3\"\n- Specific rhetorical goal stated. All 4 choices are grammatical sentences using the notes; only 1 achieves the goal.\n- Goals: emphasize difference, highlight finding, introduce topic, summarize conclusion, present cause-effect\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"rhetorical_synthesis\",\"passage\":\"...\",\"prompt\":\"The student wants to [goal]. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\nOutput ONLY a valid JSON array of EXACTLY 4 objects. No markdown. Count them before outputting.", "text": "[{\"id\":\"sat99-rw-m1-q27\",\"number\":27,\"type\":\"transitions\",\"passage\":\"The axolotl, a salamander native to lakes near Mexico City, retains its larval features—including external gills and a finned tail—throughout its adult life, a phenomenon known as neoteny. Most salamanders undergo metamorphosis, shedding these juvenile traits as they mature. The axolotl, ______ remains perpetually youthful in appearance, a trait that has made it a valuable subject for studies of regeneration and development.\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"by contrast,\",\"B\":\"for instance,\",\"C\":\"as a result,\",\"D\":\"in addition,\"},\"answer_format\":\"mcq\",\"correct_answer\":\"A\"},{\"id\":\"sat99-rw-m1-q28\",\"number\":28,\"type\":\"transitions\",\"passage\":\"Archaeologist Sarah Parcak uses satellite imagery to identify buried archaeological sites, detecting subtle variations in soil and vegetation that hint at structures beneath the surface. Traditional excavation methods can take years and require substantial funding to survey even a small region. Parcak's technique, ______ allows researchers to map hundreds of potential sites in a fraction of the time, dramatically accelerating the pace of discovery.\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"nevertheless,\",\"B\":\"likewise,\",\"C\":\"however,\",\"D\":\"for example,\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m1-q29\",\"number\":29,\"type\":\"transitions\",\"passage\":\"Economist Esther Duflo and her colleagues tested whether small financial incentives could increase childhood immunization rates in rural Rajasthan. Villages offering a kilogram of lentils per vaccination saw immunization rates rise from 6 percent to 38 percent. ______ the program proved more cost-effective than conventional outreach, because the higher participation reduced the per-child expense of operating each mobile clinic.\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"Nonetheless,\",\"B\":\"Similarly,\",\"C\":\"Moreover,\",\"D\":\"Instead,\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m1-q30\",\"number\":30,\"type\":\"transitions\",\"passage\":\"In the late 1870s, painter Mary Cassatt was invited by Edgar Degas to exhibit with the Impressionists in Paris, a group then regarded by critics as radical and unpolished. Cassatt's decision to align herself with these artists cost her the patronage of several conservative American collectors. ______ the affiliation gave her access to a circle of experimental printmakers whose techniques would reshape her own approach to composition.\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"Likewise,\",\"B\":\"In turn,\",\"C\":\"Therefore,\",\"D\":\"Still,\"},\"answer_format\":\"mcq\",\"correct_answer\":\"D\"}]" }, "status": "ok" }, { "attempts": 1, "index": 6, "output": { "costUsd": 0.121365, "model": "opus", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 3 SAT Reading & Writing questions.\nType: rhetorical_synthesis | Test: 99 | Unit: 1 | Start Q#: 31\n\nMANDATORY: Your JSON array MUST contain exactly 3 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nFollow the instructions for your type:\n\n[vocabulary_in_context]\n- 2-3 sentence passage with ______ blank from specific academic domains (molecular biology, Renaissance art, behavioral economics, astrophysics, political philosophy)\n- 4 college-level word choices from the same semantic field\n- Correct word is most PRECISE for context, not most common synonym\n- ID format: sat{testNumber}-rw-m{mod}-q{N}\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"vocabulary_in_context\",\"passage\":\"...with ______...\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[text_comprehension]\n- 3-8 sentence passage from published-quality source (journal, novel, essay) with specific names/dates/findings\n- Mix subtypes: main idea (2-3), author's purpose (1-2), inference (2-3), tone (0-1), structure (1-2)\n- Topics: natural science, social science, literature, history, arts/humanities (use each)\n- Distractors: surface-level misreading, wrong scope, opposite interpretation\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"text_comprehension\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[data_analysis]\n- Passage describing realistic data/study results requiring INTERPRETATION, not lookup\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Data patterns: diminishing returns, optimal values, conditional relationships\n- Distractors: wrong column, wrong trend direction, correlation vs causation confusion\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"data_analysis\",\"passage\":\"...\",\"visual\":{...},\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[claim_support]\n- Mix: \"Which finding supports/undermines the claim?\" AND \"Which quotation illustrates the claim?\"\n- Claims must be specific and testable. Distractors are topically related but don't address the specific mechanism.\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"claim_support\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[standard_english_conventions]\n- Passage with [underlined: text here] portion. Student picks best replacement.\n- Cover: subject-verb agreement, pronouns, verb tense, parallel structure, modifiers, commas, semicolons, apostrophes, sentence boundaries\n- \"NO CHANGE\" (choice A) correct ~25% of the time, NOT always\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"standard_english_conventions\",\"passage\":\"...with [underlined: ...]...\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[transitions]\n- Passage with ______ blank connecting two ideas. All 4 choices grammatically correct; distinction is logical.\n- Cover: contrast, addition, cause/effect, example, sequence, concession\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"transitions\",\"passage\":\"...______ ...\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[rhetorical_synthesis]\n- Notes format: \"While researching [topic], a student has taken these notes:\\n- Note 1\\n- Note 2\\n- Note 3\"\n- Specific rhetorical goal stated. All 4 choices are grammatical sentences using the notes; only 1 achieves the goal.\n- Goals: emphasize difference, highlight finding, introduce topic, summarize conclusion, present cause-effect\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"rhetorical_synthesis\",\"passage\":\"...\",\"prompt\":\"The student wants to [goal]. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\nOutput ONLY a valid JSON array of EXACTLY 3 objects. No markdown. Count them before outputting.", "text": "[{\"id\":\"sat99-rw-m1-q31\",\"number\":31,\"type\":\"rhetorical_synthesis\",\"passage\":\"While researching the migratory patterns of the Arctic tern, a student has taken these notes:\\n- The Arctic tern (Sterna paradisaea) is a small seabird with a wingspan of about 75 cm.\\n- It breeds in Arctic regions during the Northern Hemisphere summer.\\n- It winters in Antarctic waters during the Southern Hemisphere summer.\\n- A 2010 tracking study by Egevang et al. found terns travel roughly 71,000 km annually.\\n- This is the longest known migration of any animal on Earth.\",\"prompt\":\"The student wants to emphasize the remarkable nature of the Arctic tern's migration distance. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"The Arctic tern, a small seabird with a 75 cm wingspan, breeds in the Arctic and winters in Antarctic waters.\",\"B\":\"A 2010 tracking study by Egevang et al. revealed that the Arctic tern travels roughly 71,000 km annually—the longest known migration of any animal on Earth.\",\"C\":\"The Arctic tern breeds in Arctic regions during the Northern Hemisphere summer and migrates to Antarctic waters for the Southern Hemisphere summer.\",\"D\":\"Egevang and colleagues conducted a tracking study of the Arctic tern, a seabird whose scientific name is Sterna paradisaea, in 2010.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m1-q32\",\"number\":32,\"type\":\"rhetorical_synthesis\",\"passage\":\"While researching the painter Artemisia Gentileschi, a student has taken these notes:\\n- Artemisia Gentileschi (1593–1656) was an Italian Baroque painter.\\n- She trained in her father Orazio's workshop in Rome.\\n- In 1616, she became the first woman admitted to the Accademia delle Arti del Disegno in Florence.\\n- Her painting Judith Slaying Holofernes (c. 1620) is held by the Uffizi Gallery.\\n- Art historians cite the work for its dramatic chiaroscuro and unflinching depiction of violence.\",\"prompt\":\"The student wants to introduce Artemisia Gentileschi to an audience unfamiliar with her work. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"Judith Slaying Holofernes, painted around 1620, is noted by art historians for its dramatic chiaroscuro and unflinching depiction of violence.\",\"B\":\"In 1616, Artemisia Gentileschi became the first woman admitted to the Accademia delle Arti del Disegno in Florence.\",\"C\":\"Artemisia Gentileschi (1593–1656) was an Italian Baroque painter whose works, such as Judith Slaying Holofernes, are celebrated for their dramatic intensity.\",\"D\":\"Trained in her father Orazio's Roman workshop, Artemisia Gentileschi later moved to Florence, where the Uffizi Gallery now holds one of her paintings.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m1-q33\",\"number\":33,\"type\":\"rhetorical_synthesis\",\"passage\":\"While researching a behavioral economics experiment, a student has taken these notes:\\n- Researchers Dan Ariely and colleagues studied how anchoring affects willingness to pay.\\n- Participants were asked to write down the last two digits of their Social Security number.\\n- They were then asked to bid on items such as wine, chocolates, and electronics.\\n- Participants with higher digit pairs (80–99) bid up to 346% more than those with lower pairs (00–19).\\n- The digits had no logical connection to the items' actual value.\",\"prompt\":\"The student wants to highlight the central finding of the experiment. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"Dan Ariely and his colleagues conducted a behavioral economics experiment in which participants bid on items including wine, chocolates, and electronics.\",\"B\":\"Although the last two digits of a Social Security number bear no logical relation to an item's value, participants with higher digit pairs bid up to 346% more than those with lower pairs.\",\"C\":\"In one behavioral economics study, participants were asked to write down the last two digits of their Social Security number before bidding on a series of consumer items.\",\"D\":\"Anchoring is a cognitive bias in which an arbitrary number influences subsequent judgments, as Dan Ariely and colleagues demonstrated in their experiment.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"}]" }, "status": "ok" }, { "attempts": 1, "index": 7, "output": { "costUsd": 0.08304, "model": "opus", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 5 SAT Reading & Writing questions.\nType: vocabulary_in_context | Test: 99 | Unit: 2 | Start Q#: 1\n\nMANDATORY: Your JSON array MUST contain exactly 5 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nFollow the instructions for your type:\n\n[vocabulary_in_context]\n- 2-3 sentence passage with ______ blank from specific academic domains (molecular biology, Renaissance art, behavioral economics, astrophysics, political philosophy)\n- 4 college-level word choices from the same semantic field\n- Correct word is most PRECISE for context, not most common synonym\n- ID format: sat{testNumber}-rw-m{mod}-q{N}\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"vocabulary_in_context\",\"passage\":\"...with ______...\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[text_comprehension]\n- 3-8 sentence passage from published-quality source (journal, novel, essay) with specific names/dates/findings\n- Mix subtypes: main idea (2-3), author's purpose (1-2), inference (2-3), tone (0-1), structure (1-2)\n- Topics: natural science, social science, literature, history, arts/humanities (use each)\n- Distractors: surface-level misreading, wrong scope, opposite interpretation\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"text_comprehension\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[data_analysis]\n- Passage describing realistic data/study results requiring INTERPRETATION, not lookup\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Data patterns: diminishing returns, optimal values, conditional relationships\n- Distractors: wrong column, wrong trend direction, correlation vs causation confusion\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"data_analysis\",\"passage\":\"...\",\"visual\":{...},\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[claim_support]\n- Mix: \"Which finding supports/undermines the claim?\" AND \"Which quotation illustrates the claim?\"\n- Claims must be specific and testable. Distractors are topically related but don't address the specific mechanism.\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"claim_support\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[standard_english_conventions]\n- Passage with [underlined: text here] portion. Student picks best replacement.\n- Cover: subject-verb agreement, pronouns, verb tense, parallel structure, modifiers, commas, semicolons, apostrophes, sentence boundaries\n- \"NO CHANGE\" (choice A) correct ~25% of the time, NOT always\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"standard_english_conventions\",\"passage\":\"...with [underlined: ...]...\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[transitions]\n- Passage with ______ blank connecting two ideas. All 4 choices grammatically correct; distinction is logical.\n- Cover: contrast, addition, cause/effect, example, sequence, concession\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"transitions\",\"passage\":\"...______ ...\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[rhetorical_synthesis]\n- Notes format: \"While researching [topic], a student has taken these notes:\\n- Note 1\\n- Note 2\\n- Note 3\"\n- Specific rhetorical goal stated. All 4 choices are grammatical sentences using the notes; only 1 achieves the goal.\n- Goals: emphasize difference, highlight finding, introduce topic, summarize conclusion, present cause-effect\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"rhetorical_synthesis\",\"passage\":\"...\",\"prompt\":\"The student wants to [goal]. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\nOutput ONLY a valid JSON array of EXACTLY 5 objects. No markdown. Count them before outputting.", "text": "[{\"id\":\"sat99-rw-m2-q1\",\"number\":1,\"type\":\"vocabulary_in_context\",\"passage\":\"In studies of chromatin remodeling, researchers have observed that certain ATP-dependent complexes do not merely displace nucleosomes but ______ them along the DNA strand, repositioning histone octamers with remarkable spatial precision.\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"translocate\",\"B\":\"transport\",\"C\":\"transfer\",\"D\":\"transmit\"},\"answer_format\":\"mcq\",\"correct_answer\":\"A\"},{\"id\":\"sat99-rw-m2-q2\",\"number\":2,\"type\":\"vocabulary_in_context\",\"passage\":\"Art historian Patricia Rubin has argued that Vasari's biographies of Renaissance painters were never intended as neutral records; rather, they ______ a Florentine-centered narrative in which artists outside the Medici orbit were systematically diminished. This editorial shaping, she contends, distorted the historiography of Italian art for centuries.\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"disclosed\",\"B\":\"advanced\",\"C\":\"tolerated\",\"D\":\"inherited\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m2-q3\",\"number\":3,\"type\":\"vocabulary_in_context\",\"passage\":\"In behavioral economics, Richard Thaler's concept of mental accounting suggests that consumers do not treat money as fully ______. A tax refund, a birthday gift, and a paycheck of equal value will be spent quite differently, even though rational choice theory predicts identical treatment.\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"divisible\",\"B\":\"transparent\",\"C\":\"fungible\",\"D\":\"negotiable\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m2-q4\",\"number\":4,\"type\":\"vocabulary_in_context\",\"passage\":\"Observations from the James Webb Space Telescope have revealed galaxies at redshifts exceeding $z=13$ that appear far more luminous and structurally mature than existing cosmological models ______. Theorists are now reconsidering whether early star formation proceeded at rates previously thought impossible.\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"accommodate\",\"B\":\"recognize\",\"C\":\"endorse\",\"D\":\"withstand\"},\"answer_format\":\"mcq\",\"correct_answer\":\"A\"},{\"id\":\"sat99-rw-m2-q5\",\"number\":5,\"type\":\"vocabulary_in_context\",\"passage\":\"In \\\"On Liberty,\\\" John Stuart Mill warns that a citizenry may grow so accustomed to the tyranny of majority opinion that dissent becomes not merely unwelcome but nearly ______. Social pressure, he argued, can silence heterodox thought more effectively than any formal censor.\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"unthinkable\",\"B\":\"indifferent\",\"C\":\"redundant\",\"D\":\"ambivalent\"},\"answer_format\":\"mcq\",\"correct_answer\":\"A\"}]" }, "status": "ok" }, { "attempts": 1, "index": 8, "output": { "costUsd": 0.272115, "model": "opus", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 8 SAT Reading & Writing questions.\nType: text_comprehension | Test: 99 | Unit: 2 | Start Q#: 6\n\nMANDATORY: Your JSON array MUST contain exactly 8 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nFollow the instructions for your type:\n\n[vocabulary_in_context]\n- 2-3 sentence passage with ______ blank from specific academic domains (molecular biology, Renaissance art, behavioral economics, astrophysics, political philosophy)\n- 4 college-level word choices from the same semantic field\n- Correct word is most PRECISE for context, not most common synonym\n- ID format: sat{testNumber}-rw-m{mod}-q{N}\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"vocabulary_in_context\",\"passage\":\"...with ______...\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[text_comprehension]\n- 3-8 sentence passage from published-quality source (journal, novel, essay) with specific names/dates/findings\n- Mix subtypes: main idea (2-3), author's purpose (1-2), inference (2-3), tone (0-1), structure (1-2)\n- Topics: natural science, social science, literature, history, arts/humanities (use each)\n- Distractors: surface-level misreading, wrong scope, opposite interpretation\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"text_comprehension\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[data_analysis]\n- Passage describing realistic data/study results requiring INTERPRETATION, not lookup\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Data patterns: diminishing returns, optimal values, conditional relationships\n- Distractors: wrong column, wrong trend direction, correlation vs causation confusion\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"data_analysis\",\"passage\":\"...\",\"visual\":{...},\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[claim_support]\n- Mix: \"Which finding supports/undermines the claim?\" AND \"Which quotation illustrates the claim?\"\n- Claims must be specific and testable. Distractors are topically related but don't address the specific mechanism.\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"claim_support\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[standard_english_conventions]\n- Passage with [underlined: text here] portion. Student picks best replacement.\n- Cover: subject-verb agreement, pronouns, verb tense, parallel structure, modifiers, commas, semicolons, apostrophes, sentence boundaries\n- \"NO CHANGE\" (choice A) correct ~25% of the time, NOT always\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"standard_english_conventions\",\"passage\":\"...with [underlined: ...]...\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[transitions]\n- Passage with ______ blank connecting two ideas. All 4 choices grammatically correct; distinction is logical.\n- Cover: contrast, addition, cause/effect, example, sequence, concession\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"transitions\",\"passage\":\"...______ ...\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[rhetorical_synthesis]\n- Notes format: \"While researching [topic], a student has taken these notes:\\n- Note 1\\n- Note 2\\n- Note 3\"\n- Specific rhetorical goal stated. All 4 choices are grammatical sentences using the notes; only 1 achieves the goal.\n- Goals: emphasize difference, highlight finding, introduce topic, summarize conclusion, present cause-effect\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"rhetorical_synthesis\",\"passage\":\"...\",\"prompt\":\"The student wants to [goal]. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\nOutput ONLY a valid JSON array of EXACTLY 8 objects. No markdown. Count them before outputting.", "text": "[{\"id\":\"sat99-rw-m2-q6\",\"number\":6,\"type\":\"text_comprehension\",\"passage\":\"In her 2019 study of medieval manuscript illumination, art historian Elina Gertsman challenges the long-standing assumption that blank spaces in devotional texts were merely unfinished work awaiting decoration. Drawing on a corpus of over two hundred Books of Hours, Gertsman argues that voids—empty medallions, unpainted margins, and unfilled initials—were often deliberate theological statements. The absence of image, she contends, invited contemplative engagement with the unrepresentable nature of the divine, functioning as a visual corollary to apophatic theology.\",\"prompt\":\"Which choice best describes the main idea of the text?\",\"choices\":{\"A\":\"Gertsman's analysis of Books of Hours reveals that medieval illuminators frequently left manuscripts unfinished due to economic pressures from wealthy patrons.\",\"B\":\"Gertsman proposes that empty spaces in medieval devotional manuscripts should be understood as intentional theological expressions rather than incomplete craftsmanship.\",\"C\":\"Gertsman demonstrates that apophatic theology originated in the artistic practices of medieval manuscript illuminators rather than in formal philosophical treatises.\",\"D\":\"Gertsman's corpus of two hundred Books of Hours proves that medieval viewers preferred unadorned manuscripts to heavily illuminated ones.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m2-q7\",\"number\":7,\"type\":\"text_comprehension\",\"passage\":\"The following is adapted from Willa Cather's 1918 novel My Ántonia. The narrator, Jim Burden, recalls his arrival on the Nebraska prairie as a child.\\n\\nI had the feeling that the world was left behind, that we had got over the edge of it, and were outside man's jurisdiction. I had never before looked up at the sky when there was not a familiar mountain ridge against it. But this was the complete dome of heaven, all there was of it. Between that earth and that sky I felt erased, blotted out. I did not say my prayers that night: here, I felt, what would be would be.\",\"prompt\":\"Which choice best describes the function of the underlined sentence \\\"Between that earth and that sky I felt erased, blotted out\\\" in the text as a whole?\",\"choices\":{\"A\":\"It introduces a geographic description that the narrator will later contradict with more detailed observations.\",\"B\":\"It conveys the narrator's fear that the prairie's inhabitants will not welcome a newcomer from the East.\",\"C\":\"It captures a sense of self-dissolution that motivates the narrator's subsequent abandonment of a familiar ritual.\",\"D\":\"It criticizes the narrator's earlier assumption that mountains were necessary for a meaningful landscape.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m2-q8\",\"number\":8,\"type\":\"text_comprehension\",\"passage\":\"Economist Esther Duflo's randomized controlled trials in rural Rajasthan produced a counterintuitive finding: offering mothers a small bag of lentils as an incentive for vaccinating their children raised full-immunization rates from 6 percent to 39 percent, while the program's per-child cost actually fell. The lentils, Duflo observed, did not persuade skeptics of vaccine safety; rather, they solved a subtler problem. Many parents already intended to immunize but perpetually postponed the clinic visit. The modest reward supplied the nudge needed to convert latent intention into action.\",\"prompt\":\"Based on the text, what can reasonably be inferred about the parents in Duflo's study?\",\"choices\":{\"A\":\"Most parents were skeptical of vaccine safety and required material incentives to overcome their doubts.\",\"B\":\"Parents in rural Rajasthan generally placed greater value on food staples than on their children's long-term health.\",\"C\":\"Many parents' failure to vaccinate reflected procrastination rather than an objection to immunization itself.\",\"D\":\"The program's success depended primarily on reducing the travel distance between villages and immunization clinics.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m2-q9\",\"number\":9,\"type\":\"text_comprehension\",\"passage\":\"Paleontologist Jingmai O'Connor has argued that the traditional dichotomy between \\\"avian\\\" and \\\"non-avian\\\" dinosaurs obscures more than it clarifies. Her analyses of Cretaceous specimens from the Jehol Biota suggest that features once considered diagnostic of birds—feathers, wishbones, even rudimentary flight—appeared sporadically across multiple theropod lineages. To speak of \\\"the origin of birds,\\\" O'Connor writes, is to impose a retrospective clarity on what was, in the Mesozoic, a thicket of parallel experiments.\",\"prompt\":\"Which choice best describes the author's purpose in the text?\",\"choices\":{\"A\":\"To argue that O'Connor's research has definitively resolved longstanding debates about which theropod lineage gave rise to modern birds.\",\"B\":\"To explain how O'Connor's work complicates a familiar evolutionary narrative by emphasizing the distributed nature of bird-like traits.\",\"C\":\"To criticize paleontologists who continue to use the terms \\\"avian\\\" and \\\"non-avian\\\" despite evidence from the Jehol Biota.\",\"D\":\"To describe a methodological controversy about how fossil specimens from the Cretaceous should be classified and catalogued.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m2-q10\",\"number\":10,\"type\":\"text_comprehension\",\"passage\":\"In a 2021 review of urban heat-island mitigation, civil engineer Hashem Akbari surveyed three decades of rooftop albedo studies. He found that while \\\"cool roofs\\\" painted with highly reflective coatings reliably reduced indoor cooling loads by 10–20 percent in low-rise buildings, their neighborhood-scale effects were far less consistent. In some Mediterranean cities, reflected solar radiation from dense clusters of cool roofs modestly raised street-level temperatures for pedestrians, particularly in narrow canyons where the reflected light was repeatedly absorbed by adjacent walls.\",\"prompt\":\"Which choice best describes the structure of the text?\",\"choices\":{\"A\":\"It presents a widely accepted engineering solution and then describes the mechanism by which it was originally developed.\",\"B\":\"It contrasts two competing theories about urban heat mitigation and argues in favor of the more recent one.\",\"C\":\"It introduces a well-established benefit of a technology and then qualifies that benefit by noting a context-dependent drawback.\",\"D\":\"It summarizes a historical debate and concludes that the debate has been resolved by Akbari's 2021 review.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m2-q11\",\"number\":11,\"type\":\"text_comprehension\",\"passage\":\"The following is adapted from an 1891 essay by Vernon Lee on the aesthetics of music.\\n\\nI do not deny that a sonata may be dissected, its modulations charted, its thematic transformations labeled with scholarly precision. Such labor has its uses. But let us not pretend that the sum of these annotations is the sonata. The listener who emerges from the concert hall transfigured has not been moved by a catalogue of devices; she has been moved by something the catalogue cannot capture, and which the catalogue was never designed to record.\",\"prompt\":\"Which choice best describes the tone of the passage?\",\"choices\":{\"A\":\"Dismissive—Lee rejects musical analysis as a futile endeavor that produces no genuine understanding.\",\"B\":\"Measured—Lee concedes the value of technical analysis while insisting on its insufficiency for explaining musical experience.\",\"C\":\"Nostalgic—Lee laments the decline of an earlier era in which listeners were more emotionally responsive to music.\",\"D\":\"Defensive—Lee justifies her own lack of formal musical training by attacking those who possess it.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m2-q12\",\"number\":12,\"type\":\"text_comprehension\",\"passage\":\"Historians of the early Qing dynasty have long debated the significance of the 1645 queue edict, which required Han Chinese men to adopt the Manchu hairstyle on pain of death. Recent work by Joanna Waley-Cohen reframes the edict as less an assertion of ethnic supremacy than a calculated test of political loyalty. Because the hairstyle was visible at a glance, it allowed Qing officials to identify dissent without the laborious inquisitions that had undermined earlier dynasties. The edict's cruelty, Waley-Cohen suggests, should not obscure its administrative elegance.\",\"prompt\":\"Based on the text, which of the following best describes Waley-Cohen's interpretation of the queue edict?\",\"choices\":{\"A\":\"It was primarily an expression of Manchu cultural chauvinism against Han Chinese traditions.\",\"B\":\"It was a comparatively humane policy when measured against the inquisitions of earlier dynasties.\",\"C\":\"It functioned as an efficient surveillance mechanism that reduced the need for more invasive forms of investigation.\",\"D\":\"It was an impractical policy whose administrative costs outweighed its benefits to the Qing state.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m2-q13\",\"number\":13,\"type\":\"text_comprehension\",\"passage\":\"When the astronomer Vera Rubin began measuring the rotation curves of spiral galaxies in the 1970s, she expected the outer stars to orbit more slowly than those near the galactic center, just as distant planets orbit the sun more slowly than nearby ones. Instead, her spectra showed that stars at the edges of galaxies like Andromeda were moving just as fast as those much closer in. The implication, Rubin recognized, was either that the law of gravity behaved differently on galactic scales or that each galaxy was embedded in a vast halo of unseen mass.\",\"prompt\":\"Which choice best states the main idea of the text?\",\"choices\":{\"A\":\"Rubin's spectra of Andromeda disproved the Newtonian laws of gravity and established a new framework for galactic dynamics.\",\"B\":\"Rubin's observations of galactic rotation contradicted her expectations and pointed toward two possible explanations with profound implications.\",\"C\":\"Rubin's work demonstrated that spiral galaxies rotate differently from solar systems because their stars are more massive than planets.\",\"D\":\"Rubin initially misinterpreted her spectra but later corrected her findings after comparing them with planetary orbital data.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"}]" }, "status": "ok" }, { "attempts": 1, "index": 9, "output": { "costUsd": 0.17019, "model": "opus", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 4 SAT Reading & Writing questions.\nType: data_analysis | Test: 99 | Unit: 2 | Start Q#: 14\n\nMANDATORY: Your JSON array MUST contain exactly 4 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nFollow the instructions for your type:\n\n[vocabulary_in_context]\n- 2-3 sentence passage with ______ blank from specific academic domains (molecular biology, Renaissance art, behavioral economics, astrophysics, political philosophy)\n- 4 college-level word choices from the same semantic field\n- Correct word is most PRECISE for context, not most common synonym\n- ID format: sat{testNumber}-rw-m{mod}-q{N}\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"vocabulary_in_context\",\"passage\":\"...with ______...\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[text_comprehension]\n- 3-8 sentence passage from published-quality source (journal, novel, essay) with specific names/dates/findings\n- Mix subtypes: main idea (2-3), author's purpose (1-2), inference (2-3), tone (0-1), structure (1-2)\n- Topics: natural science, social science, literature, history, arts/humanities (use each)\n- Distractors: surface-level misreading, wrong scope, opposite interpretation\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"text_comprehension\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[data_analysis]\n- Passage describing realistic data/study results requiring INTERPRETATION, not lookup\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Data patterns: diminishing returns, optimal values, conditional relationships\n- Distractors: wrong column, wrong trend direction, correlation vs causation confusion\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"data_analysis\",\"passage\":\"...\",\"visual\":{...},\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[claim_support]\n- Mix: \"Which finding supports/undermines the claim?\" AND \"Which quotation illustrates the claim?\"\n- Claims must be specific and testable. Distractors are topically related but don't address the specific mechanism.\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"claim_support\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[standard_english_conventions]\n- Passage with [underlined: text here] portion. Student picks best replacement.\n- Cover: subject-verb agreement, pronouns, verb tense, parallel structure, modifiers, commas, semicolons, apostrophes, sentence boundaries\n- \"NO CHANGE\" (choice A) correct ~25% of the time, NOT always\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"standard_english_conventions\",\"passage\":\"...with [underlined: ...]...\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[transitions]\n- Passage with ______ blank connecting two ideas. All 4 choices grammatically correct; distinction is logical.\n- Cover: contrast, addition, cause/effect, example, sequence, concession\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"transitions\",\"passage\":\"...______ ...\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[rhetorical_synthesis]\n- Notes format: \"While researching [topic], a student has taken these notes:\\n- Note 1\\n- Note 2\\n- Note 3\"\n- Specific rhetorical goal stated. All 4 choices are grammatical sentences using the notes; only 1 achieves the goal.\n- Goals: emphasize difference, highlight finding, introduce topic, summarize conclusion, present cause-effect\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"rhetorical_synthesis\",\"passage\":\"...\",\"prompt\":\"The student wants to [goal]. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\nOutput ONLY a valid JSON array of EXACTLY 4 objects. No markdown. Count them before outputting.", "text": "[{\"id\":\"sat99-rw-m2-q14\",\"number\":14,\"type\":\"data_analysis\",\"passage\":\"Soil ecologist Marisol Vega investigated whether mycorrhizal fungi inoculation affects the drought tolerance of seedlings in degraded pastureland. She grew seedlings of four native tree species under identical water-restricted conditions, half with inoculation and half without, and measured survival rates after 90 days. Vega hypothesized that the benefit of inoculation would be greatest for species with naturally low drought tolerance, as measured by baseline (non-inoculated) survival.\",\"visual\":{\"type\":\"table\",\"title\":\"90-Day Seedling Survival Rates Under Drought Stress (%)\",\"headers\":[\"Species\",\"Non-inoculated\",\"Inoculated\",\"Change (pp)\"],\"rows\":[[\"Quercus rugosa\",\"72\",\"81\",\"+9\"],[\"Pinus leiophylla\",\"58\",\"74\",\"+16\"],[\"Alnus acuminata\",\"34\",\"62\",\"+28\"],[\"Prunus serotina\",\"49\",\"65\",\"+16\"]]},\"prompt\":\"Which choice best describes data from the table that support Vega's hypothesis?\",\"choices\":{\"A\":\"Alnus acuminata had the lowest non-inoculated survival rate and showed the largest increase in survival following inoculation.\",\"B\":\"Quercus rugosa had the highest non-inoculated survival rate and showed the largest increase in survival following inoculation.\",\"C\":\"Pinus leiophylla and Prunus serotina had identical increases in survival despite different baseline survival rates.\",\"D\":\"Every species showed an increase in survival following inoculation, with the average increase exceeding 15 percentage points.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"A\"},{\"id\":\"sat99-rw-m2-q15\",\"number\":15,\"type\":\"data_analysis\",\"passage\":\"In a study of consumer decision-making, economist Tarun Shah presented participants with a choice between two retirement savings plans that differed only in how contribution rates were framed: Plan X described contributions as a percentage of gross salary, while Plan Y described the same contributions as a fixed dollar amount per paycheck. Shah recorded the proportion of participants in each age group who selected Plan Y.\",\"visual\":{\"type\":\"table\",\"title\":\"Proportion Selecting Plan Y (Fixed-Dollar Framing) by Age Group\",\"headers\":[\"Age Group\",\"Plan Y Selection Rate\",\"Reported Confidence in Choice\"],\"rows\":[[\"18–29\",\"0.71\",\"Low\"],[\"30–44\",\"0.58\",\"Moderate\"],[\"45–59\",\"0.44\",\"Moderate\"],[\"60+\",\"0.29\",\"High\"]]},\"prompt\":\"Based on the table, which statement most accurately characterizes the relationship between age and framing preference?\",\"choices\":{\"A\":\"Older participants were more likely to select the fixed-dollar plan, suggesting increased susceptibility to concrete framing with age.\",\"B\":\"As age increased, preference for the fixed-dollar framing decreased, and reported confidence in the choice increased.\",\"C\":\"Confidence in the choice was lowest among participants who most frequently selected the percentage-based framing.\",\"D\":\"Age had no systematic effect on framing preference, since selection rates varied unpredictably across groups.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m2-q16\",\"number\":16,\"type\":\"data_analysis\",\"passage\":\"To assess whether urban noise affects the vocal behavior of songbirds, ornithologist Kenji Arai recorded dawn choruses of white-crowned sparrows at four sites in and around San Francisco. At each site, he measured the ambient noise level (in decibels) at 5:00 a.m. and the minimum frequency (in hertz) of the sparrows' songs, reasoning that birds in noisier environments might raise their minimum song frequency to avoid masking by low-frequency traffic noise.\",\"visual\":{\"type\":\"table\",\"title\":\"Ambient Noise and Minimum Song Frequency at Four Sites\",\"headers\":[\"Site\",\"Ambient Noise (dB)\",\"Minimum Song Frequency (Hz)\"],\"rows\":[[\"Presidio (suburban edge)\",\"42\",\"2,050\"],[\"Golden Gate Park\",\"51\",\"2,310\"],[\"Mission District\",\"63\",\"2,740\"],[\"SOMA (downtown)\",\"71\",\"3,020\"]]},\"prompt\":\"Which choice best describes data from the table that support Arai's reasoning?\",\"choices\":{\"A\":\"The Presidio site had the lowest ambient noise and the lowest minimum song frequency, while SOMA had the highest values for both.\",\"B\":\"Minimum song frequency at each site was more than 40 times the ambient noise level measured at that site.\",\"C\":\"The difference in minimum song frequency between Golden Gate Park and the Mission District was greater than the difference between the Presidio and Golden Gate Park.\",\"D\":\"Sites with higher ambient noise showed lower minimum song frequencies, contrary to what would be expected if traffic noise masked low-frequency song.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"A\"},{\"id\":\"sat99-rw-m2-q17\",\"number\":17,\"type\":\"data_analysis\",\"passage\":\"Historian Elena Costa examined the circulation numbers of four political pamphlets published in Venice between 1575 and 1590 to investigate whether length influenced popular reach. She hypothesized that shorter pamphlets would achieve wider circulation because they were cheaper to print and easier for semi-literate readers to consume in a single sitting.\",\"visual\":{\"type\":\"table\",\"title\":\"Venetian Political Pamphlets, 1575–1590\",\"headers\":[\"Pamphlet\",\"Length (pages)\",\"Estimated Copies Circulated\",\"Price (soldi)\"],\"rows\":[[\"Discorso sulla Repubblica\",\"48\",\"1,200\",\"12\"],[\"Lettera ai Cittadini\",\"8\",\"9,400\",\"2\"],[\"Trattato della Giustizia\",\"64\",\"850\",\"16\"],[\"Avviso al Popolo\",\"12\",\"7,100\",\"3\"]]},\"prompt\":\"Which choice best describes data from the table that support Costa's hypothesis?\",\"choices\":{\"A\":\"The two longest pamphlets, each exceeding 40 pages, circulated in fewer than 1,500 copies, while the two shortest, under 15 pages, each circulated in more than 7,000 copies.\",\"B\":\"Lettera ai Cittadini circulated in more copies than any other pamphlet, and its price was the lowest of the four.\",\"C\":\"The pamphlet with the highest price, Trattato della Giustizia, circulated in the fewest copies, suggesting that cost rather than length drove readership.\",\"D\":\"Avviso al Popolo was longer than Lettera ai Cittadini but circulated in fewer copies, showing that length had only a modest effect on reach.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"A\"}]" }, "status": "ok" }, { "attempts": 1, "index": 10, "output": { "costUsd": 0.145665, "model": "opus", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 4 SAT Reading & Writing questions.\nType: claim_support | Test: 99 | Unit: 2 | Start Q#: 18\n\nMANDATORY: Your JSON array MUST contain exactly 4 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nFollow the instructions for your type:\n\n[vocabulary_in_context]\n- 2-3 sentence passage with ______ blank from specific academic domains (molecular biology, Renaissance art, behavioral economics, astrophysics, political philosophy)\n- 4 college-level word choices from the same semantic field\n- Correct word is most PRECISE for context, not most common synonym\n- ID format: sat{testNumber}-rw-m{mod}-q{N}\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"vocabulary_in_context\",\"passage\":\"...with ______...\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[text_comprehension]\n- 3-8 sentence passage from published-quality source (journal, novel, essay) with specific names/dates/findings\n- Mix subtypes: main idea (2-3), author's purpose (1-2), inference (2-3), tone (0-1), structure (1-2)\n- Topics: natural science, social science, literature, history, arts/humanities (use each)\n- Distractors: surface-level misreading, wrong scope, opposite interpretation\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"text_comprehension\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[data_analysis]\n- Passage describing realistic data/study results requiring INTERPRETATION, not lookup\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Data patterns: diminishing returns, optimal values, conditional relationships\n- Distractors: wrong column, wrong trend direction, correlation vs causation confusion\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"data_analysis\",\"passage\":\"...\",\"visual\":{...},\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[claim_support]\n- Mix: \"Which finding supports/undermines the claim?\" AND \"Which quotation illustrates the claim?\"\n- Claims must be specific and testable. Distractors are topically related but don't address the specific mechanism.\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"claim_support\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[standard_english_conventions]\n- Passage with [underlined: text here] portion. Student picks best replacement.\n- Cover: subject-verb agreement, pronouns, verb tense, parallel structure, modifiers, commas, semicolons, apostrophes, sentence boundaries\n- \"NO CHANGE\" (choice A) correct ~25% of the time, NOT always\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"standard_english_conventions\",\"passage\":\"...with [underlined: ...]...\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[transitions]\n- Passage with ______ blank connecting two ideas. All 4 choices grammatically correct; distinction is logical.\n- Cover: contrast, addition, cause/effect, example, sequence, concession\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"transitions\",\"passage\":\"...______ ...\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[rhetorical_synthesis]\n- Notes format: \"While researching [topic], a student has taken these notes:\\n- Note 1\\n- Note 2\\n- Note 3\"\n- Specific rhetorical goal stated. All 4 choices are grammatical sentences using the notes; only 1 achieves the goal.\n- Goals: emphasize difference, highlight finding, introduce topic, summarize conclusion, present cause-effect\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"rhetorical_synthesis\",\"passage\":\"...\",\"prompt\":\"The student wants to [goal]. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\nOutput ONLY a valid JSON array of EXACTLY 4 objects. No markdown. Count them before outputting.", "text": "[{\"id\":\"sat99-rw-m2-q18\",\"number\":18,\"type\":\"claim_support\",\"passage\":\"In a 2019 study, marine biologist Elena Vasquez examined the foraging behavior of Caribbean reef octopuses (Octopus briareus) and proposed that these cephalopods develop individualized hunting strategies that persist across multiple foraging trips, rather than responding identically to similar environmental stimuli. Vasquez argued that this consistency reflects stable behavioral differences between individuals analogous to personality traits in vertebrates.\",\"prompt\":\"Which finding from Vasquez's study, if true, would most directly support her claim?\",\"choices\":{\"A\":\"Octopuses in the study showed a marked preference for crustacean prey over mollusks, regardless of which prey type was more abundant in their immediate habitat.\",\"B\":\"When presented with identical artificial reef environments, individual octopuses each used distinct hunting techniques—such as ambush, pursuit, or den-baiting—that each animal reliably repeated across twelve separate trials.\",\"C\":\"Octopuses observed in the wild exhibited more varied hunting behaviors than those kept in controlled laboratory conditions, suggesting that environmental complexity shapes foraging choices.\",\"D\":\"Juvenile octopuses in the study displayed a wider range of hunting techniques than mature adults, whose behaviors became increasingly uniform as they aged.\",\"E\":\"\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m2-q19\",\"number\":19,\"type\":\"claim_support\",\"passage\":\"In her 2021 analysis of Dutch Golden Age still life painting, art historian Margriet de Vries contends that Clara Peeters's signature technique of embedding tiny self-portraits within the reflective surfaces of depicted pewter vessels was not merely a display of technical virtuosity but a deliberate assertion of authorial presence in a genre that typically effaced the painter's identity. De Vries argues that these reflections functioned as visual signatures at a time when female artists were frequently denied formal attribution.\",\"prompt\":\"Which finding, if true, would most strongly support de Vries's claim?\",\"choices\":{\"A\":\"X-ray analysis of several Peeters paintings reveals that the reflected self-portraits were painted in a final, separate layer using pigments distinct from those used in the surrounding still life composition.\",\"B\":\"Peeters's contemporaries, including male still life painters such as Osias Beert, occasionally included reflections of windows or studio interiors in their depicted metalware.\",\"C\":\"Archival records indicate that several of Peeters's still lifes were sold to wealthy patrons in Antwerp without any written attribution to the artist, and that she signed far fewer works than her male contemporaries.\",\"D\":\"Technical examination of pewter vessels from the early seventeenth century shows that their polished surfaces would have produced recognizable mirror-like reflections under studio lighting.\",\"E\":\"\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m2-q20\",\"number\":20,\"type\":\"claim_support\",\"passage\":\"In The Theory of Moral Sentiments (1759), Adam Smith argues that human beings possess an innate capacity for sympathy, which he defines not as pity but as the imaginative act of placing oneself in another's situation. Smith further claims that our moral judgments of others arise primarily from this sympathetic projection rather than from abstract reasoning about duty or consequence.\",\"prompt\":\"Which quotation from The Theory of Moral Sentiments most directly illustrates Smith's claim about the origin of moral judgment?\",\"choices\":{\"A\":\"\\\"Mankind, though naturally sympathetic, never conceive, for what has befallen another, that degree of passion which naturally animates the person principally concerned.\\\"\",\"B\":\"\\\"The man of the most perfect virtue... is he who joins, to the most perfect command of his own original and selfish feelings, the most exquisite sensibility both to the original and sympathetic feelings of others.\\\"\",\"C\":\"\\\"As we have no immediate experience of what other men feel, we can form no idea of the manner in which they are affected, but by conceiving what we ourselves should feel in the like situation.\\\"\",\"D\":\"\\\"Every faculty in one man is the measure by which he judges of the like faculty in another. I judge of your sight by my sight, of your ear by my ear, of your reason by my reason.\\\"\",\"E\":\"\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m2-q21\",\"number\":21,\"type\":\"claim_support\",\"passage\":\"Behavioral economist Priya Narayanan has proposed that so-called \\\"default effects\\\"—the tendency of people to stick with pre-selected options—stem not from genuine preference alignment but from the cognitive effort required to evaluate alternatives. Narayanan predicts that when decision-makers are given more time and cognitive resources to deliberate, the influence of defaults on their final choices will diminish substantially.\",\"prompt\":\"Which finding from a controlled experiment would most directly undermine Narayanan's claim?\",\"choices\":{\"A\":\"Participants who were given a forty-five minute deliberation period and reminded to consider all options chose the default option at essentially the same rate as participants who were required to decide within thirty seconds.\",\"B\":\"Participants under time pressure were more likely to report feeling anxious about their decisions than participants given ample time to deliberate, though both groups selected the default at similar rates.\",\"C\":\"Participants who were required to justify their decisions in writing chose the default less often than those who were not asked to provide justifications, regardless of how much time they were given.\",\"D\":\"Participants who were explicitly told which option had been pre-selected as the default were more likely to choose it than participants who were not informed about the default's status.\",\"E\":\"\"},\"answer_format\":\"mcq\",\"correct_answer\":\"A\"}]" }, "status": "ok" }, { "attempts": 1, "index": 11, "output": { "costUsd": 0.07389, "model": "opus", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 5 SAT Reading & Writing questions.\nType: standard_english_conventions | Test: 99 | Unit: 2 | Start Q#: 22\n\nMANDATORY: Your JSON array MUST contain exactly 5 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nFollow the instructions for your type:\n\n[vocabulary_in_context]\n- 2-3 sentence passage with ______ blank from specific academic domains (molecular biology, Renaissance art, behavioral economics, astrophysics, political philosophy)\n- 4 college-level word choices from the same semantic field\n- Correct word is most PRECISE for context, not most common synonym\n- ID format: sat{testNumber}-rw-m{mod}-q{N}\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"vocabulary_in_context\",\"passage\":\"...with ______...\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[text_comprehension]\n- 3-8 sentence passage from published-quality source (journal, novel, essay) with specific names/dates/findings\n- Mix subtypes: main idea (2-3), author's purpose (1-2), inference (2-3), tone (0-1), structure (1-2)\n- Topics: natural science, social science, literature, history, arts/humanities (use each)\n- Distractors: surface-level misreading, wrong scope, opposite interpretation\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"text_comprehension\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[data_analysis]\n- Passage describing realistic data/study results requiring INTERPRETATION, not lookup\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Data patterns: diminishing returns, optimal values, conditional relationships\n- Distractors: wrong column, wrong trend direction, correlation vs causation confusion\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"data_analysis\",\"passage\":\"...\",\"visual\":{...},\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[claim_support]\n- Mix: \"Which finding supports/undermines the claim?\" AND \"Which quotation illustrates the claim?\"\n- Claims must be specific and testable. Distractors are topically related but don't address the specific mechanism.\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"claim_support\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[standard_english_conventions]\n- Passage with [underlined: text here] portion. Student picks best replacement.\n- Cover: subject-verb agreement, pronouns, verb tense, parallel structure, modifiers, commas, semicolons, apostrophes, sentence boundaries\n- \"NO CHANGE\" (choice A) correct ~25% of the time, NOT always\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"standard_english_conventions\",\"passage\":\"...with [underlined: ...]...\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[transitions]\n- Passage with ______ blank connecting two ideas. All 4 choices grammatically correct; distinction is logical.\n- Cover: contrast, addition, cause/effect, example, sequence, concession\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"transitions\",\"passage\":\"...______ ...\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[rhetorical_synthesis]\n- Notes format: \"While researching [topic], a student has taken these notes:\\n- Note 1\\n- Note 2\\n- Note 3\"\n- Specific rhetorical goal stated. All 4 choices are grammatical sentences using the notes; only 1 achieves the goal.\n- Goals: emphasize difference, highlight finding, introduce topic, summarize conclusion, present cause-effect\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"rhetorical_synthesis\",\"passage\":\"...\",\"prompt\":\"The student wants to [goal]. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\nOutput ONLY a valid JSON array of EXACTLY 5 objects. No markdown. Count them before outputting.", "text": "[{\"id\":\"sat99-rw-m2-q22\",\"number\":22,\"type\":\"standard_english_conventions\",\"passage\":\"The archaeologist Kathleen Kenyon, whose excavations at Jericho in the 1950s revolutionized the field's understanding of Neolithic settlement, [underlined: are] now regarded as a pioneer of stratigraphic methodology.\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"is\",\"C\":\"were\",\"D\":\"being\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m2-q23\",\"number\":23,\"type\":\"standard_english_conventions\",\"passage\":\"Marine biologists studying the Great Barrier Reef have documented a surprising resilience in certain coral species; [underlined: however these] organisms continue to reproduce even after repeated bleaching events.\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"however, these\",\"C\":\"however: these\",\"D\":\"however these,\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m2-q24\",\"number\":24,\"type\":\"standard_english_conventions\",\"passage\":\"When composing her final symphony, Florence Price drew on spirituals, European classical forms, and [underlined: she incorporated syncopated rhythms] characteristic of African American folk traditions.\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"incorporating syncopated rhythms\",\"C\":\"syncopated rhythms\",\"D\":\"syncopated rhythms were incorporated\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m2-q25\",\"number\":25,\"type\":\"standard_english_conventions\",\"passage\":\"The committee's decision to fund the observatory surprised many researchers [underlined: who's] grant proposals had been rejected in earlier rounds of review.\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"whose\",\"C\":\"whos'\",\"D\":\"whom's\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m2-q26\",\"number\":26,\"type\":\"standard_english_conventions\",\"passage\":\"By the time the expedition reached the base camp on the southern glacier, the team [underlined: has endured] three weeks of subzero temperatures and persistent whiteout conditions.\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"endures\",\"C\":\"is enduring\",\"D\":\"had endured\"},\"answer_format\":\"mcq\",\"correct_answer\":\"D\"}]" }, "status": "ok" }, { "attempts": 1, "index": 12, "output": { "costUsd": 0.082365, "model": "opus", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 4 SAT Reading & Writing questions.\nType: transitions | Test: 99 | Unit: 2 | Start Q#: 27\n\nMANDATORY: Your JSON array MUST contain exactly 4 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nFollow the instructions for your type:\n\n[vocabulary_in_context]\n- 2-3 sentence passage with ______ blank from specific academic domains (molecular biology, Renaissance art, behavioral economics, astrophysics, political philosophy)\n- 4 college-level word choices from the same semantic field\n- Correct word is most PRECISE for context, not most common synonym\n- ID format: sat{testNumber}-rw-m{mod}-q{N}\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"vocabulary_in_context\",\"passage\":\"...with ______...\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[text_comprehension]\n- 3-8 sentence passage from published-quality source (journal, novel, essay) with specific names/dates/findings\n- Mix subtypes: main idea (2-3), author's purpose (1-2), inference (2-3), tone (0-1), structure (1-2)\n- Topics: natural science, social science, literature, history, arts/humanities (use each)\n- Distractors: surface-level misreading, wrong scope, opposite interpretation\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"text_comprehension\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[data_analysis]\n- Passage describing realistic data/study results requiring INTERPRETATION, not lookup\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Data patterns: diminishing returns, optimal values, conditional relationships\n- Distractors: wrong column, wrong trend direction, correlation vs causation confusion\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"data_analysis\",\"passage\":\"...\",\"visual\":{...},\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[claim_support]\n- Mix: \"Which finding supports/undermines the claim?\" AND \"Which quotation illustrates the claim?\"\n- Claims must be specific and testable. Distractors are topically related but don't address the specific mechanism.\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"claim_support\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[standard_english_conventions]\n- Passage with [underlined: text here] portion. Student picks best replacement.\n- Cover: subject-verb agreement, pronouns, verb tense, parallel structure, modifiers, commas, semicolons, apostrophes, sentence boundaries\n- \"NO CHANGE\" (choice A) correct ~25% of the time, NOT always\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"standard_english_conventions\",\"passage\":\"...with [underlined: ...]...\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[transitions]\n- Passage with ______ blank connecting two ideas. All 4 choices grammatically correct; distinction is logical.\n- Cover: contrast, addition, cause/effect, example, sequence, concession\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"transitions\",\"passage\":\"...______ ...\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[rhetorical_synthesis]\n- Notes format: \"While researching [topic], a student has taken these notes:\\n- Note 1\\n- Note 2\\n- Note 3\"\n- Specific rhetorical goal stated. All 4 choices are grammatical sentences using the notes; only 1 achieves the goal.\n- Goals: emphasize difference, highlight finding, introduce topic, summarize conclusion, present cause-effect\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"rhetorical_synthesis\",\"passage\":\"...\",\"prompt\":\"The student wants to [goal]. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\nOutput ONLY a valid JSON array of EXACTLY 4 objects. No markdown. Count them before outputting.", "text": "[{\"id\":\"sat99-rw-m2-q27\",\"number\":27,\"type\":\"transitions\",\"passage\":\"The axolotl, a salamander native to lakes near Mexico City, retains its larval features—including external gills and a finned tail—throughout adulthood, a phenomenon known as neoteny. Most amphibians undergo metamorphosis to develop lungs and adapt to terrestrial life. ______ the axolotl remains fully aquatic, breeding and feeding in the same form it had as a juvenile.\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"By contrast,\",\"B\":\"Likewise,\",\"C\":\"As a result,\",\"D\":\"For instance,\"},\"answer_format\":\"mcq\",\"correct_answer\":\"A\"},{\"id\":\"sat99-rw-m2-q28\",\"number\":28,\"type\":\"transitions\",\"passage\":\"In her 2019 study of urban tree canopies, ecologist Mariana Velez found that neighborhoods with greater tree cover recorded summer temperatures up to 4°C lower than adjacent treeless blocks. The cooling effect persisted even during prolonged heat waves, when air-conditioning demand spiked elsewhere. ______ Velez concluded that municipal tree-planting initiatives could meaningfully reduce both energy consumption and heat-related illness in dense city centers.\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"Nevertheless,\",\"B\":\"On these grounds,\",\"C\":\"In addition,\",\"D\":\"Granted,\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m2-q29\",\"number\":29,\"type\":\"transitions\",\"passage\":\"Composer Florence Price submitted her Symphony in E minor to the 1932 Wanamaker Foundation competition and won first prize, prompting the Chicago Symphony Orchestra to perform the work the following year. For decades afterward, however, her scores were largely forgotten and her name absent from concert programs. ______ in 2009, a renovation crew discovered a trove of Price's manuscripts in her abandoned summer home, sparking renewed scholarly and performance interest.\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"Consequently,\",\"B\":\"Similarly,\",\"C\":\"Then,\",\"D\":\"Meanwhile,\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m2-q30\",\"number\":30,\"type\":\"transitions\",\"passage\":\"Behavioral economist Sendhil Mullainathan has argued that scarcity—whether of time, money, or food—captures the mind, narrowing attention to the immediate shortage and degrading performance on unrelated tasks. Critics initially dismissed the claim as overstated, suggesting that observed deficits stemmed from chronic stress rather than scarcity itself. ______ subsequent experiments in which participants were temporarily induced to feel scarce, and then released from that state, replicated the cognitive effects within a single laboratory session.\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"Moreover,\",\"B\":\"For example,\",\"C\":\"In short,\",\"D\":\"However,\"},\"answer_format\":\"mcq\",\"correct_answer\":\"D\"}]" }, "status": "ok" }, { "attempts": 1, "index": 13, "output": { "costUsd": 0.12909, "model": "opus", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 3 SAT Reading & Writing questions.\nType: rhetorical_synthesis | Test: 99 | Unit: 2 | Start Q#: 31\n\nMANDATORY: Your JSON array MUST contain exactly 3 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nFollow the instructions for your type:\n\n[vocabulary_in_context]\n- 2-3 sentence passage with ______ blank from specific academic domains (molecular biology, Renaissance art, behavioral economics, astrophysics, political philosophy)\n- 4 college-level word choices from the same semantic field\n- Correct word is most PRECISE for context, not most common synonym\n- ID format: sat{testNumber}-rw-m{mod}-q{N}\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"vocabulary_in_context\",\"passage\":\"...with ______...\",\"prompt\":\"Which choice completes the text with the most logical and precise word or phrase?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[text_comprehension]\n- 3-8 sentence passage from published-quality source (journal, novel, essay) with specific names/dates/findings\n- Mix subtypes: main idea (2-3), author's purpose (1-2), inference (2-3), tone (0-1), structure (1-2)\n- Topics: natural science, social science, literature, history, arts/humanities (use each)\n- Distractors: surface-level misreading, wrong scope, opposite interpretation\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"text_comprehension\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[data_analysis]\n- Passage describing realistic data/study results requiring INTERPRETATION, not lookup\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Data patterns: diminishing returns, optimal values, conditional relationships\n- Distractors: wrong column, wrong trend direction, correlation vs causation confusion\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"data_analysis\",\"passage\":\"...\",\"visual\":{...},\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[claim_support]\n- Mix: \"Which finding supports/undermines the claim?\" AND \"Which quotation illustrates the claim?\"\n- Claims must be specific and testable. Distractors are topically related but don't address the specific mechanism.\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"claim_support\",\"passage\":\"...\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[standard_english_conventions]\n- Passage with [underlined: text here] portion. Student picks best replacement.\n- Cover: subject-verb agreement, pronouns, verb tense, parallel structure, modifiers, commas, semicolons, apostrophes, sentence boundaries\n- \"NO CHANGE\" (choice A) correct ~25% of the time, NOT always\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"standard_english_conventions\",\"passage\":\"...with [underlined: ...]...\",\"prompt\":\"Which choice completes the text so that it conforms to the conventions of Standard English?\",\"choices\":{\"A\":\"NO CHANGE\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[transitions]\n- Passage with ______ blank connecting two ideas. All 4 choices grammatically correct; distinction is logical.\n- Cover: contrast, addition, cause/effect, example, sequence, concession\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"transitions\",\"passage\":\"...______ ...\",\"prompt\":\"Which choice completes the text with the most logical transition?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\n[rhetorical_synthesis]\n- Notes format: \"While researching [topic], a student has taken these notes:\\n- Note 1\\n- Note 2\\n- Note 3\"\n- Specific rhetorical goal stated. All 4 choices are grammatical sentences using the notes; only 1 achieves the goal.\n- Goals: emphasize difference, highlight finding, introduce topic, summarize conclusion, present cause-effect\n- Schema: {\"id\":\"...\",\"number\":N,\"type\":\"rhetorical_synthesis\",\"passage\":\"...\",\"prompt\":\"The student wants to [goal]. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\n\nOutput ONLY a valid JSON array of EXACTLY 3 objects. No markdown. Count them before outputting.", "text": "[{\"id\":\"sat99-rw-m2-q31\",\"number\":31,\"type\":\"rhetorical_synthesis\",\"passage\":\"While researching the foraging behavior of bumblebees, a student has taken the following notes:\\n- Bumblebees (genus $Bombus$) visit a wide variety of flowering plants.\\n- A 2019 study by Lars Chittka observed bumblebees in an enclosed arena with artificial flowers.\\n- Some flowers offered high-concentration sucrose solutions; others offered low-concentration solutions.\\n- Bumblebees in the study revisited high-concentration flowers significantly more often than low-concentration ones.\\n- Chittka concluded that bumblebees can distinguish between reward magnitudes and adjust foraging accordingly.\",\"prompt\":\"The student wants to present the main finding of Chittka's 2019 study. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"Bumblebees, members of the genus $Bombus$, are known to visit a wide variety of flowering plants in natural environments.\",\"B\":\"In a 2019 study, Lars Chittka observed bumblebees in an enclosed arena containing artificial flowers with varying sucrose concentrations.\",\"C\":\"In Chittka's 2019 study, bumblebees revisited artificial flowers offering high-concentration sucrose solutions significantly more often than those offering low-concentration solutions.\",\"D\":\"Chittka's research suggests that the foraging behavior of bumblebees is shaped by their ability to perceive features of their environment.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"C\"},{\"id\":\"sat99-rw-m2-q32\",\"number\":32,\"type\":\"rhetorical_synthesis\",\"passage\":\"While researching the Italian Renaissance painter Sofonisba Anguissola, a student has taken the following notes:\\n- Sofonisba Anguissola (c. 1532–1625) was a Northern Italian painter from Cremona.\\n- She trained under Bernardino Campi, an unusual arrangement for women of her era.\\n- Her 1555 painting $The$ $Chess$ $Game$ depicts her three sisters playing chess.\\n- $The$ $Chess$ $Game$ is considered an early example of a genre scene featuring identifiable individuals.\\n- Most genre scenes of the period depicted anonymous, generalized figures rather than specific people.\",\"prompt\":\"The student wants to emphasize a way in which Anguissola's $The$ $Chess$ $Game$ differed from typical genre scenes of its time. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"Sofonisba Anguissola, who lived from approximately 1532 to 1625, was a painter from the Northern Italian city of Cremona.\",\"B\":\"Whereas most genre scenes of the period depicted anonymous, generalized figures, Anguissola's 1555 painting $The$ $Chess$ $Game$ portrayed her three sisters as identifiable individuals.\",\"C\":\"Anguissola trained under the painter Bernardino Campi, an arrangement that was unusual for women during the Italian Renaissance.\",\"D\":\"Anguissola's $The$ $Chess$ $Game$, completed in 1555, depicts her three sisters engaged in a game of chess.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"B\"},{\"id\":\"sat99-rw-m2-q33\",\"number\":33,\"type\":\"rhetorical_synthesis\",\"passage\":\"While researching the formation of Saturn's rings, a student has taken the following notes:\\n- Saturn's rings are composed primarily of water ice particles ranging from micrometers to meters in size.\\n- One hypothesis proposes the rings formed roughly 4.5 billion years ago, alongside Saturn itself.\\n- A 2019 analysis of Cassini spacecraft data by Iess and colleagues estimated the rings' total mass.\\n- Based on the low mass and minimal contamination by interplanetary dust, Iess and colleagues concluded the rings are likely only 10 to 100 million years old.\\n- This estimate suggests the rings may have formed long after the planet itself, possibly from the breakup of an icy moon.\",\"prompt\":\"The student wants to introduce the topic of the research to an audience unfamiliar with Saturn's rings. Which choice most effectively uses relevant information from the notes to accomplish this goal?\",\"choices\":{\"A\":\"Saturn's rings, which are composed primarily of water ice particles varying widely in size, have been the subject of recent scientific investigation into their age and origin.\",\"B\":\"Iess and colleagues analyzed Cassini spacecraft data in 2019 to estimate the total mass of Saturn's rings.\",\"C\":\"The low mass and minimal dust contamination of Saturn's rings led Iess and colleagues to conclude that the rings are 10 to 100 million years old.\",\"D\":\"Saturn's rings may have formed from the breakup of an icy moon long after Saturn itself coalesced roughly 4.5 billion years ago.\"},\"answer_format\":\"mcq\",\"correct_answer\":\"A\"}]" }, "status": "ok" } ], "succeeded": 14 }, "gen-math": { "count": 10, "costUsd": 0, "failed": 0, "items": [ { "costUsd": 0, "model": "gemini-2.5-pro", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 7 SAT Math questions (of which 2 are student-produced-response and the rest are MCQ).\nType: algebra | Test: 99 | Unit: 1 | Start Q#: 1\n\nMANDATORY: Your JSON array MUST contain exactly 7 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nMCQ Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"algebra\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\nSPR Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"algebra\",\"prompt\":\"...\",\"answer_format\":\"student_produced\",\"correct_answer\":\"numeric_value\"}\nPlace SPR questions at the END of the array.\n\nFollow instructions for your type:\n\n[algebra] (MCQ + 2 SPR)\n- Topics: linear equations (1 var, 2 vars), systems of 2 linear equations, linear inequalities, absolute value\n- Mix word problems and abstract algebra. Use $...$ notation.\n- MCQ distractors: sign errors, arithmetic mistakes, setup errors, partial answers\n- SPR answers: clean integers or simple fractions\n\n[functions_analysis] (MCQ + 2 SPR)\n- Topics: function evaluation, domain/range, transformations (f(x)+k, f(x-h)), graphical analysis, composition f(g(x)), piecewise\n- At least one conceptual insight question (e.g., (p,q) and (q,p) on same graph)\n- Describe graphs in text. All computations yield clean results.\n\n[geometry_trigonometry] (MCQ + 2 SPR)\n- Topics: circle equations, Pythagorean theorem, special right triangles, coordinate geometry, trig ratios, volume/surface area, angle relationships\n- Optional \"visual\" field: {\"type\":\"coordinate_plane\",\"description\":\"...\"}\n- Radical answers OK (e.g., $5\\sqrt{13}$). Multi-step reasoning required.\n\n[statistics_probability] (MCQ + 2 SPR)\n- Topics: mean/median/mode, standard deviation (conceptual), probability (conditional from tables), data interpretation, scatter plots/line of best fit, sampling/bias\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Distractors: P(A|B) vs P(B|A) confusion, mean vs median, causal from correlational\n\n[advanced_math] (MCQ + 2 SPR)\n- Topics: quadratics (factoring, discriminant), polynomials (zeros, end behavior), exponentials (growth/decay), rational expressions, radical expressions, non-linear systems\n- Hardest questions on the test. Multi-step reasoning required. Elegant algebraic insight preferred.\n- All expressions must be factorable with integer coefficients. SPR answers are clean numbers.\n\nVERIFY EVERY ANSWER: Solve each problem step-by-step before finalizing correct_answer.\nOutput ONLY a valid JSON array of EXACTLY 7 objects. No markdown. Count them before outputting.", "text": "```json\n[\n {\n \"id\": \"sat99-math-m1-q1\",\n \"number\": 1,\n \"type\": \"algebra\",\n \"prompt\": \"A moving company charges a flat fee of $120 for using their truck and an additional $45 per hour for labor. If a customer's final bill was $390, how many hours of labor did the customer pay for?\",\n \"choices\": {\n \"A\": \"2.6\",\n \"B\": \"5\",\n \"C\": \"6\",\n \"D\": \"8.6\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"C\"\n },\n {\n \"id\": \"sat99-math-m1-q2\",\n \"number\": 2,\n \"type\": \"algebra\",\n \"prompt\": \"At a school fundraiser, adults' tickets were sold for $8 and students' tickets were sold for $5. A total of 150 tickets were sold, bringing in a total of $990. How many adult tickets were sold?\",\n \"choices\": {\n \"A\": \"70\",\n \"B\": \"80\",\n \"C\": \"90\",\n \"D\": \"100\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"B\"\n },\n {\n \"id\": \"sat99-math-m1-q3\",\n \"number\": 3,\n \"type\": \"algebra\",\n \"prompt\": \"The atmospheric temperature, $T$, in degrees Celsius, required for a certain scientific experiment must satisfy the inequality $|T - (-20)| \\\\le 1.5$. Which of the following represents all possible values for the temperature $T$?\",\n \"choices\": {\n \"A\": \"T \\\\ge -18.5\",\n \"B\": \"T \\\\le -21.5\",\n \"C\": \"-18.5 \\\\le T \\\\le 18.5\",\n \"D\": \"-21.5 \\\\le T \\\\le -18.5\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"D\"\n },\n {\n \"id\": \"sat99-math-m1-q4\",\n \"number\": 4,\n \"type\": \"algebra\",\n \"prompt\": \"The total cost, $C$, in dollars, for an online subscription is given by the equation $C = 15m + 20$, where $m$ is the number of months. What is the best interpretation of the number 15 in this context?\",\n \"choices\": {\n \"A\": \"The cost per month\",\n \"B\": \"The initial sign-up fee\",\n \"C\": \"The total cost for 20 months\",\n \"D\": \"The number of months included in the subscription\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"A\"\n },\n {\n \"id\": \"sat99-math-m1-q5\",\n \"number\": 5,\n \"type\": \"algebra\",\n \"prompt\": \"If $2x - 3y = 8$ and $x + y = 9$, what is the value of the expression $x - 4y$?\",\n \"choices\": {\n \"A\": \"17\",\n \"B\": \"-1\",\n \"C\": \"1\",\n \"D\": \"7\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"B\"\n },\n {\n \"id\": \"sat99-math-m1-q6\",\n \"number\": 6,\n \"type\": \"algebra\",\n \"prompt\": \"For a certain chemical process, the concentration $C$ of a substance must be maintained such that $|3C - 11| = 4$. What is one possible value for the concentration $C$?\",\n \"answer_format\": \"student_produced\",\n \"correct_answer\": \"5\"\n },\n {\n \"id\": \"sat99-math-m1-q7\",\n \"number\": 7,\n \"type\": \"algebra\",\n \"prompt\": \"If $x > 0$ and $5x - 2y = 15$ and $4x - y = 12$, what is the value of $x$?\",\n \"answer_format\": \"student_produced\",\n \"correct_answer\": \"3\"\n }\n]\n```" }, { "costUsd": 0, "model": "gemini-2.5-pro", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 5 SAT Math questions (of which 1 are student-produced-response and the rest are MCQ).\nType: functions_analysis | Test: 99 | Unit: 1 | Start Q#: 8\n\nMANDATORY: Your JSON array MUST contain exactly 5 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nMCQ Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"functions_analysis\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\nSPR Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"functions_analysis\",\"prompt\":\"...\",\"answer_format\":\"student_produced\",\"correct_answer\":\"numeric_value\"}\nPlace SPR questions at the END of the array.\n\nFollow instructions for your type:\n\n[algebra] (MCQ + 1 SPR)\n- Topics: linear equations (1 var, 2 vars), systems of 2 linear equations, linear inequalities, absolute value\n- Mix word problems and abstract algebra. Use $...$ notation.\n- MCQ distractors: sign errors, arithmetic mistakes, setup errors, partial answers\n- SPR answers: clean integers or simple fractions\n\n[functions_analysis] (MCQ + 1 SPR)\n- Topics: function evaluation, domain/range, transformations (f(x)+k, f(x-h)), graphical analysis, composition f(g(x)), piecewise\n- At least one conceptual insight question (e.g., (p,q) and (q,p) on same graph)\n- Describe graphs in text. All computations yield clean results.\n\n[geometry_trigonometry] (MCQ + 1 SPR)\n- Topics: circle equations, Pythagorean theorem, special right triangles, coordinate geometry, trig ratios, volume/surface area, angle relationships\n- Optional \"visual\" field: {\"type\":\"coordinate_plane\",\"description\":\"...\"}\n- Radical answers OK (e.g., $5\\sqrt{13}$). Multi-step reasoning required.\n\n[statistics_probability] (MCQ + 1 SPR)\n- Topics: mean/median/mode, standard deviation (conceptual), probability (conditional from tables), data interpretation, scatter plots/line of best fit, sampling/bias\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Distractors: P(A|B) vs P(B|A) confusion, mean vs median, causal from correlational\n\n[advanced_math] (MCQ + 1 SPR)\n- Topics: quadratics (factoring, discriminant), polynomials (zeros, end behavior), exponentials (growth/decay), rational expressions, radical expressions, non-linear systems\n- Hardest questions on the test. Multi-step reasoning required. Elegant algebraic insight preferred.\n- All expressions must be factorable with integer coefficients. SPR answers are clean numbers.\n\nVERIFY EVERY ANSWER: Solve each problem step-by-step before finalizing correct_answer.\nOutput ONLY a valid JSON array of EXACTLY 5 objects. No markdown. Count them before outputting.", "text": "```json\n[\n {\n \"id\": \"sat99-math-m1-q8\",\n \"number\": 8,\n \"type\": \"functions_analysis\",\n \"prompt\": \"Let the function $f$ be defined as $f(x) = x^2 + 5$ and the function $g$ be defined as $g(x) = x - 2$. What is the value of $f(g(3))$?\",\n \"choices\": {\n \"A\": \"6\",\n \"B\": \"8\",\n \"C\": \"14\",\n \"D\": \"17\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"A\"\n },\n {\n \"id\": \"sat99-math-m1-q9\",\n \"number\": 9,\n \"type\": \"functions_analysis\",\n \"prompt\": \"The graph of $y = h(x)$ is shown in the $xy$-plane. A new function, $k(x)$, is defined as $k(x) = h(x) - 2$. What is the maximum value of the function $k(x)$?\",\n \"visual\": {\n \"type\": \"coordinate_plane\",\n \"description\": \"A smooth curve representing the function h(x). The curve starts at (-4, -1), goes up to a maximum point at (-2, 5), goes down to a minimum point at (1, 0), and then goes up to the point (3, 3).\"\n },\n \"choices\": {\n \"A\": \"-2\",\n \"B\": \"3\",\n \"C\": \"5\",\n \"D\": \"7\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"B\"\n },\n {\n \"id\": \"sat99-math-m1-q10\",\n \"number\": 10,\n \"type\": \"functions_analysis\",\n \"prompt\": \"A function $f(x)$ has a domain of $x \\ge 2$ and a range of $y \\le 5$. The function $g(x)$ is defined as $g(x) = f(x+3) - 4$. Which of the following represents the domain and range of $g(x)$?\",\n \"choices\": {\n \"A\": \"Domain: $x \\ge 5$, Range: $y \\le 1$\",\n \"B\": \"Domain: $x \\ge -1$, Range: $y \\le 9$\",\n \"C\": \"Domain: $x \\ge 5$, Range: $y \\le 9$\",\n \"D\": \"Domain: $x \\ge -1$, Range: $y \\le 1$\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"D\"\n },\n {\n \"id\": \"sat99-math-m1-q11\",\n \"number\": 11,\n \"type\": \"functions_analysis\",\n \"prompt\": \"For a certain function $f$, it is known that if the point $(a, b)$ is on the graph of $y=f(x)$, then the point $(b, a)$ is also on the graph. If $f(2) = 5$, what is the value of $f(f(2))$?\",\n \"choices\": {\n \"A\": \"0\",\n \"B\": \"2\",\n \"C\": \"5\",\n \"D\": \"The value cannot be determined from the information given.\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"B\"\n },\n {\n \"id\": \"sat99-math-m1-q12\",\n \"number\": 12,\n \"type\": \"functions_analysis\",\n \"prompt\": \"For the piecewise function $g(x)$ defined below,\\n$g(x) = \\\\begin{cases} 4x-3 & \\\\text{if } x < 2 \\\\\\\\ x^2+1 & \\\\text{if } x \\\\ge 2 \\\\end{cases}$\\nWhat is the value of $g(g(1))$?\",\n \"answer_format\": \"student_produced\",\n \"correct_answer\": \"2\"\n }\n]\n```" }, { "costUsd": 0, "model": "gemini-2.5-pro", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 5 SAT Math questions (of which 1 are student-produced-response and the rest are MCQ).\nType: geometry_trigonometry | Test: 99 | Unit: 1 | Start Q#: 13\n\nMANDATORY: Your JSON array MUST contain exactly 5 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nMCQ Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"geometry_trigonometry\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\nSPR Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"geometry_trigonometry\",\"prompt\":\"...\",\"answer_format\":\"student_produced\",\"correct_answer\":\"numeric_value\"}\nPlace SPR questions at the END of the array.\n\nFollow instructions for your type:\n\n[algebra] (MCQ + 1 SPR)\n- Topics: linear equations (1 var, 2 vars), systems of 2 linear equations, linear inequalities, absolute value\n- Mix word problems and abstract algebra. Use $...$ notation.\n- MCQ distractors: sign errors, arithmetic mistakes, setup errors, partial answers\n- SPR answers: clean integers or simple fractions\n\n[functions_analysis] (MCQ + 1 SPR)\n- Topics: function evaluation, domain/range, transformations (f(x)+k, f(x-h)), graphical analysis, composition f(g(x)), piecewise\n- At least one conceptual insight question (e.g., (p,q) and (q,p) on same graph)\n- Describe graphs in text. All computations yield clean results.\n\n[geometry_trigonometry] (MCQ + 1 SPR)\n- Topics: circle equations, Pythagorean theorem, special right triangles, coordinate geometry, trig ratios, volume/surface area, angle relationships\n- Optional \"visual\" field: {\"type\":\"coordinate_plane\",\"description\":\"...\"}\n- Radical answers OK (e.g., $5\\sqrt{13}$). Multi-step reasoning required.\n\n[statistics_probability] (MCQ + 1 SPR)\n- Topics: mean/median/mode, standard deviation (conceptual), probability (conditional from tables), data interpretation, scatter plots/line of best fit, sampling/bias\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Distractors: P(A|B) vs P(B|A) confusion, mean vs median, causal from correlational\n\n[advanced_math] (MCQ + 1 SPR)\n- Topics: quadratics (factoring, discriminant), polynomials (zeros, end behavior), exponentials (growth/decay), rational expressions, radical expressions, non-linear systems\n- Hardest questions on the test. Multi-step reasoning required. Elegant algebraic insight preferred.\n- All expressions must be factorable with integer coefficients. SPR answers are clean numbers.\n\nVERIFY EVERY ANSWER: Solve each problem step-by-step before finalizing correct_answer.\nOutput ONLY a valid JSON array of EXACTLY 5 objects. No markdown. Count them before outputting.", "text": "```json\n[\n {\n \"id\": \"sat99-math-m1-q13\",\n \"number\": 13,\n \"type\": \"geometry_trigonometry\",\n \"prompt\": \"The equation of a circle in the $xy$-plane is $x^2 + y^2 - 6x + 8y + 21 = 0$. What is the radius of the circle?\",\n \"choices\": {\n \"A\": \"5\",\n \"B\": \"$\\\\sqrt{21}$\",\n \"C\": \"2\",\n \"D\": \"4\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"C\"\n },\n {\n \"id\": \"sat99-math-m1-q14\",\n \"number\": 14,\n \"type\": \"geometry_trigonometry\",\n \"prompt\": \"In the $xy$-plane, a line segment is drawn from the origin to a point $P$ in Quadrant I. If the length of the segment is 8 and it makes an angle $\\\\theta$ with the positive $x$-axis such that $\\\\tan\\\\theta = 1$, what is the $y$-coordinate of point $P$?\",\n \"choices\": {\n \"A\": \"4\",\n \"B\": \"$4\\\\sqrt{2}$\",\n \"C\": \"8\",\n \"D\": \"$8\\\\sqrt{2}$\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"B\"\n },\n {\n \"id\": \"sat99-math-m1-q15\",\n \"number\": 15,\n \"type\": \"geometry_trigonometry\",\n \"prompt\": \"In right triangle $XYZ$, the measure of angle $Z$ is $90^{\\\\circ}$. If $\\\\sin(X) = \\\\frac{7}{25}$, what is the value of $\\\\cos(Y)$?\",\n \"choices\": {\n \"A\": \"$\\\\frac{24}{25}$\",\n \"B\": \"$\\\\frac{7}{24}$\",\n \"C\": \"$\\\\frac{25}{7}$\",\n \"D\": \"$\\\\frac{7}{25}$\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"D\"\n },\n {\n \"id\": \"sat99-math-m1-q16\",\n \"number\": 16,\n \"type\": \"geometry_trigonometry\",\n \"prompt\": \"A right circular cylinder is inscribed within a cube. The volume of the cube is 64 cubic centimeters. What is the volume of the cylinder, in cubic centimeters?\",\n \"choices\": {\n \"A\": \"$8\\\\pi$\",\n \"B\": \"$16\\\\pi$\",\n \"C\": \"$32\\\\pi$\",\n \"D\": \"$64\\\\pi$\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"B\"\n },\n {\n \"id\": \"sat99-math-m1-q17\",\n \"number\": 17,\n \"type\": \"geometry_trigonometry\",\n \"prompt\": \"A circle has a radius of 10. The area of a sector of the circle is $10\\\\pi$. What is the measure of the sector's central angle, in degrees?\",\n \"answer_format\": \"student_produced\",\n \"correct_answer\": \"36\"\n }\n]\n```" }, { "costUsd": 0, "model": "gemini-2.5-pro", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 5 SAT Math questions (of which 1 are student-produced-response and the rest are MCQ).\nType: statistics_probability | Test: 99 | Unit: 1 | Start Q#: 18\n\nMANDATORY: Your JSON array MUST contain exactly 5 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nMCQ Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"statistics_probability\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\nSPR Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"statistics_probability\",\"prompt\":\"...\",\"answer_format\":\"student_produced\",\"correct_answer\":\"numeric_value\"}\nPlace SPR questions at the END of the array.\n\nFollow instructions for your type:\n\n[algebra] (MCQ + 1 SPR)\n- Topics: linear equations (1 var, 2 vars), systems of 2 linear equations, linear inequalities, absolute value\n- Mix word problems and abstract algebra. Use $...$ notation.\n- MCQ distractors: sign errors, arithmetic mistakes, setup errors, partial answers\n- SPR answers: clean integers or simple fractions\n\n[functions_analysis] (MCQ + 1 SPR)\n- Topics: function evaluation, domain/range, transformations (f(x)+k, f(x-h)), graphical analysis, composition f(g(x)), piecewise\n- At least one conceptual insight question (e.g., (p,q) and (q,p) on same graph)\n- Describe graphs in text. All computations yield clean results.\n\n[geometry_trigonometry] (MCQ + 1 SPR)\n- Topics: circle equations, Pythagorean theorem, special right triangles, coordinate geometry, trig ratios, volume/surface area, angle relationships\n- Optional \"visual\" field: {\"type\":\"coordinate_plane\",\"description\":\"...\"}\n- Radical answers OK (e.g., $5\\sqrt{13}$). Multi-step reasoning required.\n\n[statistics_probability] (MCQ + 1 SPR)\n- Topics: mean/median/mode, standard deviation (conceptual), probability (conditional from tables), data interpretation, scatter plots/line of best fit, sampling/bias\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Distractors: P(A|B) vs P(B|A) confusion, mean vs median, causal from correlational\n\n[advanced_math] (MCQ + 1 SPR)\n- Topics: quadratics (factoring, discriminant), polynomials (zeros, end behavior), exponentials (growth/decay), rational expressions, radical expressions, non-linear systems\n- Hardest questions on the test. Multi-step reasoning required. Elegant algebraic insight preferred.\n- All expressions must be factorable with integer coefficients. SPR answers are clean numbers.\n\nVERIFY EVERY ANSWER: Solve each problem step-by-step before finalizing correct_answer.\nOutput ONLY a valid JSON array of EXACTLY 5 objects. No markdown. Count them before outputting.", "text": "```json\n[\n {\n \"id\": \"sat99-math-m1-q18\",\n \"number\": 18,\n \"type\": \"statistics_probability\",\n \"prompt\": \"A scatterplot shows the relationship between the number of hours a student practices a musical instrument per week and the number of mistakes made during a performance. The line of best fit is given by the equation $y = -2.6x + 15$, where $y$ is the predicted number of mistakes and $x$ is the number of hours practiced. What is the best interpretation of the slope of the line of best fit?\",\n \"choices\": {\n \"A\": \"For each additional hour of practice, the predicted number of mistakes decreases by 2.6.\",\n \"B\": \"A student who does not practice is predicted to make 15 mistakes.\",\n \"C\": \"For every 2.6 additional mistakes made, the predicted number of practice hours decreases by 1.\",\n \"D\": \"The maximum possible number of mistakes is 15.\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"A\"\n },\n {\n \"id\": \"sat99-math-m1-q19\",\n \"number\": 19,\n \"type\": \"statistics_probability\",\n \"visual\": {\n \"type\": \"table\",\n \"title\": \"Participation in After-School Programs\",\n \"headers\": [\"\", \"Sports\", \"Arts\", \"Total\"],\n \"rows\": [\n [\"Juniors\", 85, 65, 150],\n [\"Seniors\", 95, 55, 150],\n [\"Total\", 180, 120, 300]\n ]\n },\n \"prompt\": \"The table above shows the distribution of 300 juniors and seniors at a high school based on their participation in after-school programs. If a student who participates in sports is selected at random, what is the probability that the student is a senior?\",\n \"choices\": {\n \"A\": \"$\\\\frac{95}{300}$\",\n \"B\": \"$\\\\frac{150}{300}$\",\n \"C\": \"$\\\\frac{95}{150}$\",\n \"D\": \"$\\\\frac{95}{180}$\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"D\"\n },\n {\n \"id\": \"sat99-math-m1-q20\",\n \"number\": 20,\n \"type\": \"statistics_probability\",\n \"prompt\": \"A data set consists of 9 integers. The numbers are 10, 12, 12, 13, 15, 17, 19, 20, and 45. Which of the following statements is true about the mean, median, and mode of this data set?\",\n \"choices\": {\n \"A\": \"The mean is greater than the median.\",\n \"B\": \"The median is greater than the mean.\",\n \"C\": \"The mode is greater than the mean.\",\n \"D\": \"The mean, median, and mode are all equal.\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"A\"\n },\n {\n \"id\": \"sat99-math-m1-q21\",\n \"number\": 21,\n \"type\": \"statistics_probability\",\n \"prompt\": \"An environmental scientist wants to estimate the proportion of trees in a large forest that are infected with a certain fungus. She divides the forest into 100 equally sized plots, randomly selects 10 plots, and examines every tree within those 10 plots. Which of the following best describes this sampling method?\",\n \"choices\": {\n \"A\": \"Simple random sampling\",\n \"B\": \"Cluster sampling\",\n \"C\": \"Stratified random sampling\",\n \"D\": \"Convenience sampling\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"B\"\n },\n {\n \"id\": \"sat99-math-m1-q22\",\n \"number\": 22,\n \"type\": \"statistics_probability\",\n \"visual\": {\n \"type\": \"table\",\n \"title\": \"Vehicle Types at a Service Center\",\n \"headers\": [\"\", \"Requires Oil Change\", \"Requires Tire Rotation\", \"Total\"],\n \"rows\": [\n [\"Sedan\", 50, 30, 80],\n [\"SUV\", 40, 35, 75],\n [\"Total\", 90, 65, 155]\n ]\n },\n \"prompt\": \"The table above shows the types of vehicles and the primary service requested at an auto service center on one day. Based on the table, what proportion of sedans required an oil change? (Express your answer as a decimal or fraction).\",\n \"answer_format\": \"student_produced\",\n \"correct_answer\": \"5/8\"\n }\n]\n```" }, { "costUsd": 0, "model": "gemini-2.5-pro", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 5 SAT Math questions (of which 2 are student-produced-response and the rest are MCQ).\nType: advanced_math | Test: 99 | Unit: 1 | Start Q#: 23\n\nMANDATORY: Your JSON array MUST contain exactly 5 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nMCQ Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"advanced_math\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\nSPR Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"advanced_math\",\"prompt\":\"...\",\"answer_format\":\"student_produced\",\"correct_answer\":\"numeric_value\"}\nPlace SPR questions at the END of the array.\n\nFollow instructions for your type:\n\n[algebra] (MCQ + 2 SPR)\n- Topics: linear equations (1 var, 2 vars), systems of 2 linear equations, linear inequalities, absolute value\n- Mix word problems and abstract algebra. Use $...$ notation.\n- MCQ distractors: sign errors, arithmetic mistakes, setup errors, partial answers\n- SPR answers: clean integers or simple fractions\n\n[functions_analysis] (MCQ + 2 SPR)\n- Topics: function evaluation, domain/range, transformations (f(x)+k, f(x-h)), graphical analysis, composition f(g(x)), piecewise\n- At least one conceptual insight question (e.g., (p,q) and (q,p) on same graph)\n- Describe graphs in text. All computations yield clean results.\n\n[geometry_trigonometry] (MCQ + 2 SPR)\n- Topics: circle equations, Pythagorean theorem, special right triangles, coordinate geometry, trig ratios, volume/surface area, angle relationships\n- Optional \"visual\" field: {\"type\":\"coordinate_plane\",\"description\":\"...\"}\n- Radical answers OK (e.g., $5\\sqrt{13}$). Multi-step reasoning required.\n\n[statistics_probability] (MCQ + 2 SPR)\n- Topics: mean/median/mode, standard deviation (conceptual), probability (conditional from tables), data interpretation, scatter plots/line of best fit, sampling/bias\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Distractors: P(A|B) vs P(B|A) confusion, mean vs median, causal from correlational\n\n[advanced_math] (MCQ + 2 SPR)\n- Topics: quadratics (factoring, discriminant), polynomials (zeros, end behavior), exponentials (growth/decay), rational expressions, radical expressions, non-linear systems\n- Hardest questions on the test. Multi-step reasoning required. Elegant algebraic insight preferred.\n- All expressions must be factorable with integer coefficients. SPR answers are clean numbers.\n\nVERIFY EVERY ANSWER: Solve each problem step-by-step before finalizing correct_answer.\nOutput ONLY a valid JSON array of EXACTLY 5 objects. No markdown. Count them before outputting.", "text": "```json\n[\n {\n \"id\": \"sat99-math-m1-q23\",\n \"number\": 23,\n \"type\": \"advanced_math\",\n \"prompt\": \"The quadratic equation $kx^2 - 6x + 3 = 0$, where $k$ is a constant, has exactly one real solution. What is the value of $k$?\",\n \"choices\": {\n \"A\": \"-3\",\n \"B\": \"0\",\n \"C\": \"3\",\n \"D\": \"9\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"C\"\n },\n {\n \"id\": \"sat99-math-m1-q24\",\n \"number\": 24,\n \"type\": \"advanced_math\",\n \"prompt\": \"The polynomial $P(x) = x^3 - 2x^2 - 5x + 6$ has a zero at $x=1$. What is the sum of the other two zeros of the polynomial?\",\n \"choices\": {\n \"A\": \"-1\",\n \"B\": \"1\",\n \"C\": \"2\",\n \"D\": \"5\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"B\"\n },\n {\n \"id\": \"sat99-math-m1-q25\",\n \"number\": 25,\n \"type\": \"advanced_math\",\n \"prompt\": \"If $x > 2$, which of the following expressions is equivalent to $\\\\frac{x - \\\\frac{4}{x}}{1 + \\\\frac{2}{x}}$?\",\n \"choices\": {\n \"A\": \"x - 2\",\n \"B\": \"x + 2\",\n \"C\": \"\\\\frac{x-4}{3}\",\n \"D\": \"x - 4\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"A\"\n },\n {\n \"id\": \"sat99-math-m1-q26\",\n \"number\": 26,\n \"type\": \"advanced_math\",\n \"prompt\": \"A system of two equations is given by $y = x^2 - 5x + 7$ and $y - x = 2$. If $(x, y)$ is a solution to the system, what is one possible value for $x$?\",\n \"answer_format\": \"student_produced\",\n \"correct_answer\": \"5\"\n },\n {\n \"id\": \"sat99-math-m1-q27\",\n \"number\": 27,\n \"type\": \"advanced_math\",\n \"prompt\": \"If $n > 0$ and $\\\\sqrt[3]{n^2} \\\\cdot n^k = n^{5/3}$, what is the value of $k$?\",\n \"answer_format\": \"student_produced\",\n \"correct_answer\": \"1\"\n }\n]\n```" }, { "costUsd": 0, "model": "gemini-2.5-pro", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 7 SAT Math questions (of which 2 are student-produced-response and the rest are MCQ).\nType: algebra | Test: 99 | Unit: 2 | Start Q#: 1\n\nMANDATORY: Your JSON array MUST contain exactly 7 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nMCQ Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"algebra\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\nSPR Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"algebra\",\"prompt\":\"...\",\"answer_format\":\"student_produced\",\"correct_answer\":\"numeric_value\"}\nPlace SPR questions at the END of the array.\n\nFollow instructions for your type:\n\n[algebra] (MCQ + 2 SPR)\n- Topics: linear equations (1 var, 2 vars), systems of 2 linear equations, linear inequalities, absolute value\n- Mix word problems and abstract algebra. Use $...$ notation.\n- MCQ distractors: sign errors, arithmetic mistakes, setup errors, partial answers\n- SPR answers: clean integers or simple fractions\n\n[functions_analysis] (MCQ + 2 SPR)\n- Topics: function evaluation, domain/range, transformations (f(x)+k, f(x-h)), graphical analysis, composition f(g(x)), piecewise\n- At least one conceptual insight question (e.g., (p,q) and (q,p) on same graph)\n- Describe graphs in text. All computations yield clean results.\n\n[geometry_trigonometry] (MCQ + 2 SPR)\n- Topics: circle equations, Pythagorean theorem, special right triangles, coordinate geometry, trig ratios, volume/surface area, angle relationships\n- Optional \"visual\" field: {\"type\":\"coordinate_plane\",\"description\":\"...\"}\n- Radical answers OK (e.g., $5\\sqrt{13}$). Multi-step reasoning required.\n\n[statistics_probability] (MCQ + 2 SPR)\n- Topics: mean/median/mode, standard deviation (conceptual), probability (conditional from tables), data interpretation, scatter plots/line of best fit, sampling/bias\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Distractors: P(A|B) vs P(B|A) confusion, mean vs median, causal from correlational\n\n[advanced_math] (MCQ + 2 SPR)\n- Topics: quadratics (factoring, discriminant), polynomials (zeros, end behavior), exponentials (growth/decay), rational expressions, radical expressions, non-linear systems\n- Hardest questions on the test. Multi-step reasoning required. Elegant algebraic insight preferred.\n- All expressions must be factorable with integer coefficients. SPR answers are clean numbers.\n\nVERIFY EVERY ANSWER: Solve each problem step-by-step before finalizing correct_answer.\nOutput ONLY a valid JSON array of EXACTLY 7 objects. No markdown. Count them before outputting.", "text": "```json\n[\n {\n \"id\": \"sat99-math-m2-q1\",\n \"number\": 1,\n \"type\": \"algebra\",\n \"prompt\": \"A car rental agency charges a flat daily fee of $40 plus an additional charge of $0.15 per mile driven. If a customer's total bill for a one-day rental was $98.50, how many miles did the customer drive?\",\n \"choices\": {\n \"A\": \"267\",\n \"B\": \"39\",\n \"C\": \"390\",\n \"D\": \"657\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"C\"\n },\n {\n \"id\": \"sat99-math-m2-q2\",\n \"number\": 2,\n \"type\": \"algebra\",\n \"prompt\": \"If $|2x - 3| = 11$, what is one possible value of $x$?\",\n \"choices\": {\n \"A\": \"7\",\n \"B\": \"4\",\n \"C\": \"-7\",\n \"D\": \"-4\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"D\"\n },\n {\n \"id\": \"sat99-math-m2-q3\",\n \"number\": 3,\n \"type\": \"algebra\",\n \"prompt\": \"At a food truck, 2 tacos and 1 burrito cost a total of $11.00. At the same food truck, 1 taco and 3 burritos cost a total of $18.00. What is the cost of one burrito?\",\n \"choices\": {\n \"A\": \"$2.50\",\n \"B\": \"$3.00\",\n \"C\": \"$5.50\",\n \"D\": \"$6.00\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"C\"\n },\n {\n \"id\": \"sat99-math-m2-q4\",\n \"number\": 4,\n \"type\": \"algebra\",\n \"prompt\": \"Which of the following values of $x$ is a solution to the inequality $8 - 3x < -7$?\",\n \"choices\": {\n \"A\": \"6\",\n \"B\": \"5\",\n \"C\": \"4\",\n \"D\": \"-5\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"A\"\n },\n {\n \"id\": \"sat99-math-m2-q5\",\n \"number\": 5,\n \"type\": \"algebra\",\n \"prompt\": \"A small business's weekly profit, $P$, in dollars, is modeled by the function $P(x) = 150x - 2700$, where $x$ represents the number of items sold. What is the minimum number of items the business must sell in a week to realize a profit?\",\n \"choices\": {\n \"A\": \"19\",\n \"B\": \"18\",\n \"C\": \"17\",\n \"D\": \"16\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"A\"\n },\n {\n \"id\": \"sat99-math-m2-q6\",\n \"number\": 6,\n \"type\": \"algebra\",\n \"prompt\": \"If $3x + 2y = 19$ and $x - y = 3$, what is the value of $x$?\",\n \"answer_format\": \"student_produced\",\n \"correct_answer\": \"5\"\n },\n {\n \"id\": \"sat99-math-m2-q7\",\n \"number\": 7,\n \"type\": \"algebra\",\n \"prompt\": \"If $\\\\frac{2}{3}(x - 4) = \\\\frac{1}{2}(x + 2)$, what is the value of $x$?\",\n \"answer_format\": \"student_produced\",\n \"correct_answer\": \"22\"\n }\n]\n```" }, { "costUsd": 0, "model": "gemini-2.5-pro", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 5 SAT Math questions (of which 1 are student-produced-response and the rest are MCQ).\nType: functions_analysis | Test: 99 | Unit: 2 | Start Q#: 8\n\nMANDATORY: Your JSON array MUST contain exactly 5 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nMCQ Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"functions_analysis\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\nSPR Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"functions_analysis\",\"prompt\":\"...\",\"answer_format\":\"student_produced\",\"correct_answer\":\"numeric_value\"}\nPlace SPR questions at the END of the array.\n\nFollow instructions for your type:\n\n[algebra] (MCQ + 1 SPR)\n- Topics: linear equations (1 var, 2 vars), systems of 2 linear equations, linear inequalities, absolute value\n- Mix word problems and abstract algebra. Use $...$ notation.\n- MCQ distractors: sign errors, arithmetic mistakes, setup errors, partial answers\n- SPR answers: clean integers or simple fractions\n\n[functions_analysis] (MCQ + 1 SPR)\n- Topics: function evaluation, domain/range, transformations (f(x)+k, f(x-h)), graphical analysis, composition f(g(x)), piecewise\n- At least one conceptual insight question (e.g., (p,q) and (q,p) on same graph)\n- Describe graphs in text. All computations yield clean results.\n\n[geometry_trigonometry] (MCQ + 1 SPR)\n- Topics: circle equations, Pythagorean theorem, special right triangles, coordinate geometry, trig ratios, volume/surface area, angle relationships\n- Optional \"visual\" field: {\"type\":\"coordinate_plane\",\"description\":\"...\"}\n- Radical answers OK (e.g., $5\\sqrt{13}$). Multi-step reasoning required.\n\n[statistics_probability] (MCQ + 1 SPR)\n- Topics: mean/median/mode, standard deviation (conceptual), probability (conditional from tables), data interpretation, scatter plots/line of best fit, sampling/bias\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Distractors: P(A|B) vs P(B|A) confusion, mean vs median, causal from correlational\n\n[advanced_math] (MCQ + 1 SPR)\n- Topics: quadratics (factoring, discriminant), polynomials (zeros, end behavior), exponentials (growth/decay), rational expressions, radical expressions, non-linear systems\n- Hardest questions on the test. Multi-step reasoning required. Elegant algebraic insight preferred.\n- All expressions must be factorable with integer coefficients. SPR answers are clean numbers.\n\nVERIFY EVERY ANSWER: Solve each problem step-by-step before finalizing correct_answer.\nOutput ONLY a valid JSON array of EXACTLY 5 objects. No markdown. Count them before outputting.", "text": "```json\n[\n {\n \"id\": \"sat99-math-m2-q8\",\n \"number\": 8,\n \"type\": \"functions_analysis\",\n \"prompt\": \"For the function $f$ defined by $f(x) = 2x^2 - 5x + 1$, what is the value of $f(3) - f(-1)$?\",\n \"choices\": {\n \"A\": \"-4\",\n \"B\": \"-1\",\n \"C\": \"4\",\n \"D\": \"12\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"A\"\n },\n {\n \"id\": \"sat99-math-m2-q9\",\n \"number\": 9,\n \"type\": \"functions_analysis\",\n \"prompt\": \"The graph of a function $g$ is shown in the $xy$-plane. The function $h$ is defined by $h(x) = g(x-2) + 3$. If the point $(-1, 4)$ is on the graph of $g$, what is a point on the graph of $h$?\",\n \"choices\": {\n \"A\": \"(-3, 7)\",\n \"B\": \"(-3, 1)\",\n \"C\": \"(1, 7)\",\n \"D\": \"(1, 1)\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"C\"\n },\n {\n \"id\": \"sat99-math-m2-q10\",\n \"number\": 10,\n \"type\": \"functions_analysis\",\n \"prompt\": \"The function $f$ is an even function, meaning $f(x) = f(-x)$ for all values of $x$. The function $g$ is an odd function, meaning $g(-x) = -g(x)$ for all values of $x$. If $f(2) = 5$ and $g(-3) = -4$, what is the value of $f(-2) + g(3)$?\",\n \"choices\": {\n \"A\": \"-9\",\n \"B\": \"1\",\n \"C\": \"9\",\n \"D\": \"The value cannot be determined from the information given.\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"C\"\n },\n {\n \"id\": \"sat99-math-m2-q11\",\n \"number\": 11,\n \"type\": \"functions_analysis\",\n \"prompt\": \"Let the functions $p$ and $q$ be defined by $p(x) = x^2 - 4$ and $q(x) = 2x + 1$. For what positive value of $x$ is $p(q(x)) = 0$?\",\n \"choices\": {\n \"A\": \"$\\\\frac{1}{2}$\",\n \"B\": \"$\\\\frac{3}{2}$\",\n \"C\": \"$2$\",\n \"D\": \"$5$\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"A\"\n },\n {\n \"id\": \"sat99-math-m2-q12\",\n \"number\": 12,\n \"type\": \"functions_analysis\",\n \"prompt\": \"The function $f$ is defined below.\\n$f(x) = \\\\begin{cases} x^2+3 & \\\\text{if } x < 1 \\\\\\\\ 2x-1 & \\\\text{if } x \\\\ge 1 \\\\end{cases}$\\nWhat is the value of $f(f(0))$?\",\n \"answer_format\": \"student_produced\",\n \"correct_answer\": \"5\"\n }\n]\n```" }, { "costUsd": 0, "model": "gemini-2.5-pro", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 5 SAT Math questions (of which 1 are student-produced-response and the rest are MCQ).\nType: geometry_trigonometry | Test: 99 | Unit: 2 | Start Q#: 13\n\nMANDATORY: Your JSON array MUST contain exactly 5 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nMCQ Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"geometry_trigonometry\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\nSPR Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"geometry_trigonometry\",\"prompt\":\"...\",\"answer_format\":\"student_produced\",\"correct_answer\":\"numeric_value\"}\nPlace SPR questions at the END of the array.\n\nFollow instructions for your type:\n\n[algebra] (MCQ + 1 SPR)\n- Topics: linear equations (1 var, 2 vars), systems of 2 linear equations, linear inequalities, absolute value\n- Mix word problems and abstract algebra. Use $...$ notation.\n- MCQ distractors: sign errors, arithmetic mistakes, setup errors, partial answers\n- SPR answers: clean integers or simple fractions\n\n[functions_analysis] (MCQ + 1 SPR)\n- Topics: function evaluation, domain/range, transformations (f(x)+k, f(x-h)), graphical analysis, composition f(g(x)), piecewise\n- At least one conceptual insight question (e.g., (p,q) and (q,p) on same graph)\n- Describe graphs in text. All computations yield clean results.\n\n[geometry_trigonometry] (MCQ + 1 SPR)\n- Topics: circle equations, Pythagorean theorem, special right triangles, coordinate geometry, trig ratios, volume/surface area, angle relationships\n- Optional \"visual\" field: {\"type\":\"coordinate_plane\",\"description\":\"...\"}\n- Radical answers OK (e.g., $5\\sqrt{13}$). Multi-step reasoning required.\n\n[statistics_probability] (MCQ + 1 SPR)\n- Topics: mean/median/mode, standard deviation (conceptual), probability (conditional from tables), data interpretation, scatter plots/line of best fit, sampling/bias\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Distractors: P(A|B) vs P(B|A) confusion, mean vs median, causal from correlational\n\n[advanced_math] (MCQ + 1 SPR)\n- Topics: quadratics (factoring, discriminant), polynomials (zeros, end behavior), exponentials (growth/decay), rational expressions, radical expressions, non-linear systems\n- Hardest questions on the test. Multi-step reasoning required. Elegant algebraic insight preferred.\n- All expressions must be factorable with integer coefficients. SPR answers are clean numbers.\n\nVERIFY EVERY ANSWER: Solve each problem step-by-step before finalizing correct_answer.\nOutput ONLY a valid JSON array of EXACTLY 5 objects. No markdown. Count them before outputting.", "text": "```json\n[\n {\n \"id\": \"sat99-math-m2-q13\",\n \"number\": 13,\n \"type\": \"geometry_trigonometry\",\n \"prompt\": \"A circle in the $xy$-plane has its center at $(3, -5)$. The point $(8, 7)$ lies on the circle. Which of the following is the equation of the circle?\",\n \"choices\": {\n \"A\": \"$(x-3)^2 + (y+5)^2 = 13$\",\n \"B\": \"$(x-8)^2 + (y-7)^2 = 169$\",\n \"C\": \"$(x+3)^2 + (y-5)^2 = 169$\",\n \"D\": \"$(x-3)^2 + (y+5)^2 = 169$\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"D\"\n },\n {\n \"id\": \"sat99-math-m2-q14\",\n \"number\": 14,\n \"type\": \"geometry_trigonometry\",\n \"prompt\": \"In right triangle $ABC$, the measure of angle $C$ is $90^{\\\\circ}$. If $\\\\sin(A) = \\\\frac{8}{17}$, what is the value of $\\\\cos(B)$?\",\n \"choices\": {\n \"A\": \"$\\\\frac{15}{17}$\",\n \"B\": \"$\\\\frac{8}{15}$\",\n \"C\": \"$\\\\frac{8}{17}$\",\n \"D\": \"$\\\\frac{17}{8}$\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"C\"\n },\n {\n \"id\": \"sat99-math-m2-q15\",\n \"number\": 15,\n \"type\": \"geometry_trigonometry\",\n \"prompt\": \"A cylindrical grain silo has a radius of 10 feet and a height of 40 feet. The top of the silo is a hemisphere with the same radius. What is the total volume of the silo in cubic feet?\",\n \"choices\": {\n \"A\": \"$4000\\\\pi$\",\n \"B\": \"$\\\\frac{14000}{3}\\\\pi$\",\n \"C\": \"$\\\\frac{16000}{3}\\\\pi$\",\n \"D\": \"$4400\\\\pi$\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"B\"\n },\n {\n \"id\": \"sat99-math-m2-q16\",\n \"number\": 16,\n \"type\": \"geometry_trigonometry\",\n \"prompt\": \"In the $xy$-plane, line $L$ is defined by the equation $y = -\\\\frac{1}{3}x + 5$. Line $M$ is perpendicular to line $L$ and passes through the point $(-2, 2)$. What is the $y$-intercept of line $M$?\",\n \"choices\": {\n \"A\": \"8\",\n \"B\": \"-8\",\n \"C\": \"3\",\n \"D\": \"$-\\\\frac{1}{3}$\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"A\"\n },\n {\n \"id\": \"sat99-math-m2-q17\",\n \"number\": 17,\n \"type\": \"geometry_trigonometry\",\n \"prompt\": \"A ladder is placed against a vertical wall. The base of the ladder is 7 feet away from the base of the wall, and the ladder makes an angle of $60^{\\\\circ}$ with the ground. What is the length of the ladder in feet?\",\n \"answer_format\": \"student_produced\",\n \"correct_answer\": \"14\"\n }\n]\n```" }, { "costUsd": 0, "model": "gemini-2.5-pro", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 5 SAT Math questions (of which 1 are student-produced-response and the rest are MCQ).\nType: statistics_probability | Test: 99 | Unit: 2 | Start Q#: 18\n\nMANDATORY: Your JSON array MUST contain exactly 5 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nMCQ Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"statistics_probability\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\nSPR Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"statistics_probability\",\"prompt\":\"...\",\"answer_format\":\"student_produced\",\"correct_answer\":\"numeric_value\"}\nPlace SPR questions at the END of the array.\n\nFollow instructions for your type:\n\n[algebra] (MCQ + 1 SPR)\n- Topics: linear equations (1 var, 2 vars), systems of 2 linear equations, linear inequalities, absolute value\n- Mix word problems and abstract algebra. Use $...$ notation.\n- MCQ distractors: sign errors, arithmetic mistakes, setup errors, partial answers\n- SPR answers: clean integers or simple fractions\n\n[functions_analysis] (MCQ + 1 SPR)\n- Topics: function evaluation, domain/range, transformations (f(x)+k, f(x-h)), graphical analysis, composition f(g(x)), piecewise\n- At least one conceptual insight question (e.g., (p,q) and (q,p) on same graph)\n- Describe graphs in text. All computations yield clean results.\n\n[geometry_trigonometry] (MCQ + 1 SPR)\n- Topics: circle equations, Pythagorean theorem, special right triangles, coordinate geometry, trig ratios, volume/surface area, angle relationships\n- Optional \"visual\" field: {\"type\":\"coordinate_plane\",\"description\":\"...\"}\n- Radical answers OK (e.g., $5\\sqrt{13}$). Multi-step reasoning required.\n\n[statistics_probability] (MCQ + 1 SPR)\n- Topics: mean/median/mode, standard deviation (conceptual), probability (conditional from tables), data interpretation, scatter plots/line of best fit, sampling/bias\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Distractors: P(A|B) vs P(B|A) confusion, mean vs median, causal from correlational\n\n[advanced_math] (MCQ + 1 SPR)\n- Topics: quadratics (factoring, discriminant), polynomials (zeros, end behavior), exponentials (growth/decay), rational expressions, radical expressions, non-linear systems\n- Hardest questions on the test. Multi-step reasoning required. Elegant algebraic insight preferred.\n- All expressions must be factorable with integer coefficients. SPR answers are clean numbers.\n\nVERIFY EVERY ANSWER: Solve each problem step-by-step before finalizing correct_answer.\nOutput ONLY a valid JSON array of EXACTLY 5 objects. No markdown. Count them before outputting.", "text": "```json\n[\n {\n \"id\": \"sat99-math-m2-q18\",\n \"number\": 18,\n \"type\": \"statistics_probability\",\n \"prompt\": \"A market research firm surveyed 200 owners of either a sedan or an SUV. Each owner was asked whether their next vehicle purchase would be a new or used vehicle. The results are partially summarized in the table below. Of those surveyed, 60% owned a sedan.\",\n \"visual\": {\n \"type\": \"table\",\n \"title\": \"Next Vehicle Purchase Plan\",\n \"headers\": [\"Current Vehicle\", \"New\", \"Used\", \"Total\"],\n \"rows\": [\n [\"Sedan\", 72, \"\", \"\"],\n [\"SUV\", \"\", \"\", \"\"],\n [\"Total\", 110, 90, 200]\n ]\n },\n \"prompt_continue\": \"Based on the table, what is the probability that a person who currently owns an SUV will plan to purchase a new vehicle?\",\n \"choices\": {\n \"A\": \"0.34\",\n \"B\": \"0.475\",\n \"C\": \"0.525\",\n \"D\": \"0.65\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"B\"\n },\n {\n \"id\": \"sat99-math-m2-q19\",\n \"number\": 19,\n \"type\": \"statistics_probability\",\n \"prompt\": \"The dot plot above shows the number of goals scored by a soccer team in 15 games. If the game with the highest number of goals is removed from the data, which of the following statements about the mean and median of the resulting data is true?\",\n \"visual\": {\n \"type\": \"dot_plot\",\n \"description\": \"A dot plot titled 'Goals Scored in 15 Games'. The x-axis is labeled 'Number of Goals'. There are 3 dots at 0, 5 dots at 1, 4 dots at 2, 2 dots at 3, and 1 dot at 6.\"\n },\n \"choices\": {\n \"A\": \"The mean will decrease, and the median will remain the same.\",\n \"B\": \"The mean will decrease, and the median will decrease.\",\n \"C\": \"The mean will remain the same, and the median will decrease.\",\n \"D\": \"Both the mean and the median will remain the same.\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"A\"\n },\n {\n \"id\": \"sat99-math-m2-q20\",\n \"number\": 20,\n \"type\": \"statistics_probability\",\n \"prompt\": \"A sample of 50 trees was selected from a large forest. The height, in feet, and the trunk diameter, in inches, was measured for each tree. The data are shown in the scatterplot, along with a line of best fit. The equation of the line is $h = 3.8d + 15.2$, where $h$ is the height and $d$ is the trunk diameter.\",\n \"visual\": {\n \"type\": \"scatter_plot\",\n \"description\": \"A scatterplot titled 'Tree Height vs. Trunk Diameter' showing a positive linear association. The x-axis is 'Trunk Diameter (inches)' from 5 to 25. The y-axis is 'Height (feet)' from 20 to 120. A line of best fit with a positive slope is drawn through the data points.\"\n },\n \"prompt_continue\": \"Which of the following is the best interpretation of the slope of the line of best fit?\",\n \"choices\": {\n \"A\": \"For a tree with a trunk diameter of 0 inches, the predicted height is 3.8 feet.\",\n \"B\": \"For every 1-foot increase in a tree's height, the predicted trunk diameter increases by 3.8 inches.\",\n \"C\": \"For every 3.8-inch increase in a tree's trunk diameter, the predicted height increases by 1 foot.\",\n \"D\": \"For every 1-inch increase in a tree's trunk diameter, the predicted height increases by 3.8 feet.\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"D\"\n },\n {\n \"id\": \"sat99-math-m2-q21\",\n \"number\": 21,\n \"type\": \"statistics_probability\",\n \"prompt\": \"The test scores for two history classes are summarized by the box plots below.\",\n \"visual\": {\n \"type\": \"box_plot\",\n \"description\": \"Two box plots on a single axis from 50 to 100. 'Class A': min=55, Q1=65, median=75, Q3=80, max=90. 'Class B': min=60, Q1=75, median=80, Q3=85, max=90. The box for Class A is wider than the box for Class B, indicating a larger interquartile range.\"\n },\n \"prompt_continue\": \"Based on the box plots, which of the following must be true?\",\n \"choices\": {\n \"A\": \"The mean score of Class A is greater than the mean score of Class B.\",\n \"B\": \"The range of scores for Class A is greater than the range of scores for Class B.\",\n \"C\": \"The median score of Class B is greater than the median score of Class A.\",\n \"D\": \"There are more students in Class B than in Class A.\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"C\"\n },\n {\n \"id\": \"sat99-math-m2-q22\",\n \"number\": 22,\n \"type\": \"statistics_probability\",\n \"prompt\": \"A dataset consists of 7 distinct integers. The median of the dataset is 12, the minimum value is 5, and the range is 14. What is the maximum possible value of the mean of the dataset, rounded to the nearest tenth?\",\n \"answer_format\": \"student_produced\",\n \"correct_answer\": \"14.3\"\n }\n]\n```" }, { "costUsd": 0, "model": "gemini-2.5-pro", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 5 SAT Math questions (of which 2 are student-produced-response and the rest are MCQ).\nType: advanced_math | Test: 99 | Unit: 2 | Start Q#: 23\n\nMANDATORY: Your JSON array MUST contain exactly 5 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nMCQ Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"advanced_math\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\nSPR Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"advanced_math\",\"prompt\":\"...\",\"answer_format\":\"student_produced\",\"correct_answer\":\"numeric_value\"}\nPlace SPR questions at the END of the array.\n\nFollow instructions for your type:\n\n[algebra] (MCQ + 2 SPR)\n- Topics: linear equations (1 var, 2 vars), systems of 2 linear equations, linear inequalities, absolute value\n- Mix word problems and abstract algebra. Use $...$ notation.\n- MCQ distractors: sign errors, arithmetic mistakes, setup errors, partial answers\n- SPR answers: clean integers or simple fractions\n\n[functions_analysis] (MCQ + 2 SPR)\n- Topics: function evaluation, domain/range, transformations (f(x)+k, f(x-h)), graphical analysis, composition f(g(x)), piecewise\n- At least one conceptual insight question (e.g., (p,q) and (q,p) on same graph)\n- Describe graphs in text. All computations yield clean results.\n\n[geometry_trigonometry] (MCQ + 2 SPR)\n- Topics: circle equations, Pythagorean theorem, special right triangles, coordinate geometry, trig ratios, volume/surface area, angle relationships\n- Optional \"visual\" field: {\"type\":\"coordinate_plane\",\"description\":\"...\"}\n- Radical answers OK (e.g., $5\\sqrt{13}$). Multi-step reasoning required.\n\n[statistics_probability] (MCQ + 2 SPR)\n- Topics: mean/median/mode, standard deviation (conceptual), probability (conditional from tables), data interpretation, scatter plots/line of best fit, sampling/bias\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Distractors: P(A|B) vs P(B|A) confusion, mean vs median, causal from correlational\n\n[advanced_math] (MCQ + 2 SPR)\n- Topics: quadratics (factoring, discriminant), polynomials (zeros, end behavior), exponentials (growth/decay), rational expressions, radical expressions, non-linear systems\n- Hardest questions on the test. Multi-step reasoning required. Elegant algebraic insight preferred.\n- All expressions must be factorable with integer coefficients. SPR answers are clean numbers.\n\nVERIFY EVERY ANSWER: Solve each problem step-by-step before finalizing correct_answer.\nOutput ONLY a valid JSON array of EXACTLY 5 objects. No markdown. Count them before outputting.", "text": "```json\n[\n {\n \"id\": \"sat99-math-m2-q23\",\n \"number\": 23,\n \"type\": \"advanced_math\",\n \"prompt\": \"The system of equations $y = x^2 - 4x + 7$ and $y = 2x + 2$ is graphed in the $xy$-plane. What is the sum of the $y$-coordinates of the two intersection points?\",\n \"choices\": {\n \"A\": \"6\",\n \"B\": \"12\",\n \"C\": \"18\",\n \"D\": \"24\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"C\"\n },\n {\n \"id\": \"sat99-math-m2-q24\",\n \"number\": 24,\n \"type\": \"advanced_math\",\n \"prompt\": \"A polynomial function $p(x)$ has zeros at $x=-1$ and $x=4$, where the zero at $x=4$ has a multiplicity of 2. The graph of $y=p(x)$ has a positive $y$-intercept. Which of the following could be the equation for $p(x)$?\",\n \"choices\": {\n \"A\": \"p(x) = (x+1)(x-4)^2\",\n \"B\": \"p(x) = (x-1)(x+4)^2\",\n \"C\": \"p(x) = -(x-1)(x+4)^2\",\n \"D\": \"p(x) = -(x+1)(x-4)^2\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"D\"\n },\n {\n \"id\": \"sat99-math-m2-q25\",\n \"number\": 25,\n \"type\": \"advanced_math\",\n \"prompt\": \"The function $f$ is defined by $f(x) = \\\\frac{2x^2 + 7x - 15}{x+5}$. The graph of $y=f(x)$ is a line with a point of discontinuity. What is the $y$-coordinate of the point of discontinuity?\",\n \"choices\": {\n \"A\": \"-13\",\n \"B\": \"-5\",\n \"C\": \"2\",\n \"D\": \"3\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"A\"\n },\n {\n \"id\": \"sat99-math-m2-q26\",\n \"number\": 26,\n \"type\": \"advanced_math\",\n \"prompt\": \"If $x > 0$, what value of $x$ is the solution to the equation $\\\\sqrt{2x+3} = x$?\",\n \"answer_format\": \"student_produced\",\n \"correct_answer\": \"3\"\n },\n {\n \"id\": \"sat99-math-m2-q27\",\n \"number\": 27,\n \"type\": \"advanced_math\",\n \"prompt\": \"The value of an investment is modeled by the function $V(t) = 100(1.5)^{\\\\frac{t}{4}}$, where $V(t)$ is the value of the investment in thousands of dollars and $t$ is the number of years since the initial investment was made. After how many years will the value of the investment be $225,000?\",\n \"answer_format\": \"student_produced\",\n \"correct_answer\": \"8\"\n }\n]\n```" } ], "results": [ { "attempts": 1, "index": 0, "output": { "costUsd": 0, "model": "gemini-2.5-pro", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 7 SAT Math questions (of which 2 are student-produced-response and the rest are MCQ).\nType: algebra | Test: 99 | Unit: 1 | Start Q#: 1\n\nMANDATORY: Your JSON array MUST contain exactly 7 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nMCQ Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"algebra\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\nSPR Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"algebra\",\"prompt\":\"...\",\"answer_format\":\"student_produced\",\"correct_answer\":\"numeric_value\"}\nPlace SPR questions at the END of the array.\n\nFollow instructions for your type:\n\n[algebra] (MCQ + 2 SPR)\n- Topics: linear equations (1 var, 2 vars), systems of 2 linear equations, linear inequalities, absolute value\n- Mix word problems and abstract algebra. Use $...$ notation.\n- MCQ distractors: sign errors, arithmetic mistakes, setup errors, partial answers\n- SPR answers: clean integers or simple fractions\n\n[functions_analysis] (MCQ + 2 SPR)\n- Topics: function evaluation, domain/range, transformations (f(x)+k, f(x-h)), graphical analysis, composition f(g(x)), piecewise\n- At least one conceptual insight question (e.g., (p,q) and (q,p) on same graph)\n- Describe graphs in text. All computations yield clean results.\n\n[geometry_trigonometry] (MCQ + 2 SPR)\n- Topics: circle equations, Pythagorean theorem, special right triangles, coordinate geometry, trig ratios, volume/surface area, angle relationships\n- Optional \"visual\" field: {\"type\":\"coordinate_plane\",\"description\":\"...\"}\n- Radical answers OK (e.g., $5\\sqrt{13}$). Multi-step reasoning required.\n\n[statistics_probability] (MCQ + 2 SPR)\n- Topics: mean/median/mode, standard deviation (conceptual), probability (conditional from tables), data interpretation, scatter plots/line of best fit, sampling/bias\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Distractors: P(A|B) vs P(B|A) confusion, mean vs median, causal from correlational\n\n[advanced_math] (MCQ + 2 SPR)\n- Topics: quadratics (factoring, discriminant), polynomials (zeros, end behavior), exponentials (growth/decay), rational expressions, radical expressions, non-linear systems\n- Hardest questions on the test. Multi-step reasoning required. Elegant algebraic insight preferred.\n- All expressions must be factorable with integer coefficients. SPR answers are clean numbers.\n\nVERIFY EVERY ANSWER: Solve each problem step-by-step before finalizing correct_answer.\nOutput ONLY a valid JSON array of EXACTLY 7 objects. No markdown. Count them before outputting.", "text": "```json\n[\n {\n \"id\": \"sat99-math-m1-q1\",\n \"number\": 1,\n \"type\": \"algebra\",\n \"prompt\": \"A moving company charges a flat fee of $120 for using their truck and an additional $45 per hour for labor. If a customer's final bill was $390, how many hours of labor did the customer pay for?\",\n \"choices\": {\n \"A\": \"2.6\",\n \"B\": \"5\",\n \"C\": \"6\",\n \"D\": \"8.6\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"C\"\n },\n {\n \"id\": \"sat99-math-m1-q2\",\n \"number\": 2,\n \"type\": \"algebra\",\n \"prompt\": \"At a school fundraiser, adults' tickets were sold for $8 and students' tickets were sold for $5. A total of 150 tickets were sold, bringing in a total of $990. How many adult tickets were sold?\",\n \"choices\": {\n \"A\": \"70\",\n \"B\": \"80\",\n \"C\": \"90\",\n \"D\": \"100\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"B\"\n },\n {\n \"id\": \"sat99-math-m1-q3\",\n \"number\": 3,\n \"type\": \"algebra\",\n \"prompt\": \"The atmospheric temperature, $T$, in degrees Celsius, required for a certain scientific experiment must satisfy the inequality $|T - (-20)| \\\\le 1.5$. Which of the following represents all possible values for the temperature $T$?\",\n \"choices\": {\n \"A\": \"T \\\\ge -18.5\",\n \"B\": \"T \\\\le -21.5\",\n \"C\": \"-18.5 \\\\le T \\\\le 18.5\",\n \"D\": \"-21.5 \\\\le T \\\\le -18.5\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"D\"\n },\n {\n \"id\": \"sat99-math-m1-q4\",\n \"number\": 4,\n \"type\": \"algebra\",\n \"prompt\": \"The total cost, $C$, in dollars, for an online subscription is given by the equation $C = 15m + 20$, where $m$ is the number of months. What is the best interpretation of the number 15 in this context?\",\n \"choices\": {\n \"A\": \"The cost per month\",\n \"B\": \"The initial sign-up fee\",\n \"C\": \"The total cost for 20 months\",\n \"D\": \"The number of months included in the subscription\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"A\"\n },\n {\n \"id\": \"sat99-math-m1-q5\",\n \"number\": 5,\n \"type\": \"algebra\",\n \"prompt\": \"If $2x - 3y = 8$ and $x + y = 9$, what is the value of the expression $x - 4y$?\",\n \"choices\": {\n \"A\": \"17\",\n \"B\": \"-1\",\n \"C\": \"1\",\n \"D\": \"7\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"B\"\n },\n {\n \"id\": \"sat99-math-m1-q6\",\n \"number\": 6,\n \"type\": \"algebra\",\n \"prompt\": \"For a certain chemical process, the concentration $C$ of a substance must be maintained such that $|3C - 11| = 4$. What is one possible value for the concentration $C$?\",\n \"answer_format\": \"student_produced\",\n \"correct_answer\": \"5\"\n },\n {\n \"id\": \"sat99-math-m1-q7\",\n \"number\": 7,\n \"type\": \"algebra\",\n \"prompt\": \"If $x > 0$ and $5x - 2y = 15$ and $4x - y = 12$, what is the value of $x$?\",\n \"answer_format\": \"student_produced\",\n \"correct_answer\": \"3\"\n }\n]\n```" }, "status": "ok" }, { "attempts": 1, "index": 1, "output": { "costUsd": 0, "model": "gemini-2.5-pro", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 5 SAT Math questions (of which 1 are student-produced-response and the rest are MCQ).\nType: functions_analysis | Test: 99 | Unit: 1 | Start Q#: 8\n\nMANDATORY: Your JSON array MUST contain exactly 5 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nMCQ Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"functions_analysis\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\nSPR Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"functions_analysis\",\"prompt\":\"...\",\"answer_format\":\"student_produced\",\"correct_answer\":\"numeric_value\"}\nPlace SPR questions at the END of the array.\n\nFollow instructions for your type:\n\n[algebra] (MCQ + 1 SPR)\n- Topics: linear equations (1 var, 2 vars), systems of 2 linear equations, linear inequalities, absolute value\n- Mix word problems and abstract algebra. Use $...$ notation.\n- MCQ distractors: sign errors, arithmetic mistakes, setup errors, partial answers\n- SPR answers: clean integers or simple fractions\n\n[functions_analysis] (MCQ + 1 SPR)\n- Topics: function evaluation, domain/range, transformations (f(x)+k, f(x-h)), graphical analysis, composition f(g(x)), piecewise\n- At least one conceptual insight question (e.g., (p,q) and (q,p) on same graph)\n- Describe graphs in text. All computations yield clean results.\n\n[geometry_trigonometry] (MCQ + 1 SPR)\n- Topics: circle equations, Pythagorean theorem, special right triangles, coordinate geometry, trig ratios, volume/surface area, angle relationships\n- Optional \"visual\" field: {\"type\":\"coordinate_plane\",\"description\":\"...\"}\n- Radical answers OK (e.g., $5\\sqrt{13}$). Multi-step reasoning required.\n\n[statistics_probability] (MCQ + 1 SPR)\n- Topics: mean/median/mode, standard deviation (conceptual), probability (conditional from tables), data interpretation, scatter plots/line of best fit, sampling/bias\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Distractors: P(A|B) vs P(B|A) confusion, mean vs median, causal from correlational\n\n[advanced_math] (MCQ + 1 SPR)\n- Topics: quadratics (factoring, discriminant), polynomials (zeros, end behavior), exponentials (growth/decay), rational expressions, radical expressions, non-linear systems\n- Hardest questions on the test. Multi-step reasoning required. Elegant algebraic insight preferred.\n- All expressions must be factorable with integer coefficients. SPR answers are clean numbers.\n\nVERIFY EVERY ANSWER: Solve each problem step-by-step before finalizing correct_answer.\nOutput ONLY a valid JSON array of EXACTLY 5 objects. No markdown. Count them before outputting.", "text": "```json\n[\n {\n \"id\": \"sat99-math-m1-q8\",\n \"number\": 8,\n \"type\": \"functions_analysis\",\n \"prompt\": \"Let the function $f$ be defined as $f(x) = x^2 + 5$ and the function $g$ be defined as $g(x) = x - 2$. What is the value of $f(g(3))$?\",\n \"choices\": {\n \"A\": \"6\",\n \"B\": \"8\",\n \"C\": \"14\",\n \"D\": \"17\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"A\"\n },\n {\n \"id\": \"sat99-math-m1-q9\",\n \"number\": 9,\n \"type\": \"functions_analysis\",\n \"prompt\": \"The graph of $y = h(x)$ is shown in the $xy$-plane. A new function, $k(x)$, is defined as $k(x) = h(x) - 2$. What is the maximum value of the function $k(x)$?\",\n \"visual\": {\n \"type\": \"coordinate_plane\",\n \"description\": \"A smooth curve representing the function h(x). The curve starts at (-4, -1), goes up to a maximum point at (-2, 5), goes down to a minimum point at (1, 0), and then goes up to the point (3, 3).\"\n },\n \"choices\": {\n \"A\": \"-2\",\n \"B\": \"3\",\n \"C\": \"5\",\n \"D\": \"7\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"B\"\n },\n {\n \"id\": \"sat99-math-m1-q10\",\n \"number\": 10,\n \"type\": \"functions_analysis\",\n \"prompt\": \"A function $f(x)$ has a domain of $x \\ge 2$ and a range of $y \\le 5$. The function $g(x)$ is defined as $g(x) = f(x+3) - 4$. Which of the following represents the domain and range of $g(x)$?\",\n \"choices\": {\n \"A\": \"Domain: $x \\ge 5$, Range: $y \\le 1$\",\n \"B\": \"Domain: $x \\ge -1$, Range: $y \\le 9$\",\n \"C\": \"Domain: $x \\ge 5$, Range: $y \\le 9$\",\n \"D\": \"Domain: $x \\ge -1$, Range: $y \\le 1$\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"D\"\n },\n {\n \"id\": \"sat99-math-m1-q11\",\n \"number\": 11,\n \"type\": \"functions_analysis\",\n \"prompt\": \"For a certain function $f$, it is known that if the point $(a, b)$ is on the graph of $y=f(x)$, then the point $(b, a)$ is also on the graph. If $f(2) = 5$, what is the value of $f(f(2))$?\",\n \"choices\": {\n \"A\": \"0\",\n \"B\": \"2\",\n \"C\": \"5\",\n \"D\": \"The value cannot be determined from the information given.\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"B\"\n },\n {\n \"id\": \"sat99-math-m1-q12\",\n \"number\": 12,\n \"type\": \"functions_analysis\",\n \"prompt\": \"For the piecewise function $g(x)$ defined below,\\n$g(x) = \\\\begin{cases} 4x-3 & \\\\text{if } x < 2 \\\\\\\\ x^2+1 & \\\\text{if } x \\\\ge 2 \\\\end{cases}$\\nWhat is the value of $g(g(1))$?\",\n \"answer_format\": \"student_produced\",\n \"correct_answer\": \"2\"\n }\n]\n```" }, "status": "ok" }, { "attempts": 1, "index": 2, "output": { "costUsd": 0, "model": "gemini-2.5-pro", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 5 SAT Math questions (of which 1 are student-produced-response and the rest are MCQ).\nType: geometry_trigonometry | Test: 99 | Unit: 1 | Start Q#: 13\n\nMANDATORY: Your JSON array MUST contain exactly 5 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nMCQ Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"geometry_trigonometry\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\nSPR Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"geometry_trigonometry\",\"prompt\":\"...\",\"answer_format\":\"student_produced\",\"correct_answer\":\"numeric_value\"}\nPlace SPR questions at the END of the array.\n\nFollow instructions for your type:\n\n[algebra] (MCQ + 1 SPR)\n- Topics: linear equations (1 var, 2 vars), systems of 2 linear equations, linear inequalities, absolute value\n- Mix word problems and abstract algebra. Use $...$ notation.\n- MCQ distractors: sign errors, arithmetic mistakes, setup errors, partial answers\n- SPR answers: clean integers or simple fractions\n\n[functions_analysis] (MCQ + 1 SPR)\n- Topics: function evaluation, domain/range, transformations (f(x)+k, f(x-h)), graphical analysis, composition f(g(x)), piecewise\n- At least one conceptual insight question (e.g., (p,q) and (q,p) on same graph)\n- Describe graphs in text. All computations yield clean results.\n\n[geometry_trigonometry] (MCQ + 1 SPR)\n- Topics: circle equations, Pythagorean theorem, special right triangles, coordinate geometry, trig ratios, volume/surface area, angle relationships\n- Optional \"visual\" field: {\"type\":\"coordinate_plane\",\"description\":\"...\"}\n- Radical answers OK (e.g., $5\\sqrt{13}$). Multi-step reasoning required.\n\n[statistics_probability] (MCQ + 1 SPR)\n- Topics: mean/median/mode, standard deviation (conceptual), probability (conditional from tables), data interpretation, scatter plots/line of best fit, sampling/bias\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Distractors: P(A|B) vs P(B|A) confusion, mean vs median, causal from correlational\n\n[advanced_math] (MCQ + 1 SPR)\n- Topics: quadratics (factoring, discriminant), polynomials (zeros, end behavior), exponentials (growth/decay), rational expressions, radical expressions, non-linear systems\n- Hardest questions on the test. Multi-step reasoning required. Elegant algebraic insight preferred.\n- All expressions must be factorable with integer coefficients. SPR answers are clean numbers.\n\nVERIFY EVERY ANSWER: Solve each problem step-by-step before finalizing correct_answer.\nOutput ONLY a valid JSON array of EXACTLY 5 objects. No markdown. Count them before outputting.", "text": "```json\n[\n {\n \"id\": \"sat99-math-m1-q13\",\n \"number\": 13,\n \"type\": \"geometry_trigonometry\",\n \"prompt\": \"The equation of a circle in the $xy$-plane is $x^2 + y^2 - 6x + 8y + 21 = 0$. What is the radius of the circle?\",\n \"choices\": {\n \"A\": \"5\",\n \"B\": \"$\\\\sqrt{21}$\",\n \"C\": \"2\",\n \"D\": \"4\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"C\"\n },\n {\n \"id\": \"sat99-math-m1-q14\",\n \"number\": 14,\n \"type\": \"geometry_trigonometry\",\n \"prompt\": \"In the $xy$-plane, a line segment is drawn from the origin to a point $P$ in Quadrant I. If the length of the segment is 8 and it makes an angle $\\\\theta$ with the positive $x$-axis such that $\\\\tan\\\\theta = 1$, what is the $y$-coordinate of point $P$?\",\n \"choices\": {\n \"A\": \"4\",\n \"B\": \"$4\\\\sqrt{2}$\",\n \"C\": \"8\",\n \"D\": \"$8\\\\sqrt{2}$\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"B\"\n },\n {\n \"id\": \"sat99-math-m1-q15\",\n \"number\": 15,\n \"type\": \"geometry_trigonometry\",\n \"prompt\": \"In right triangle $XYZ$, the measure of angle $Z$ is $90^{\\\\circ}$. If $\\\\sin(X) = \\\\frac{7}{25}$, what is the value of $\\\\cos(Y)$?\",\n \"choices\": {\n \"A\": \"$\\\\frac{24}{25}$\",\n \"B\": \"$\\\\frac{7}{24}$\",\n \"C\": \"$\\\\frac{25}{7}$\",\n \"D\": \"$\\\\frac{7}{25}$\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"D\"\n },\n {\n \"id\": \"sat99-math-m1-q16\",\n \"number\": 16,\n \"type\": \"geometry_trigonometry\",\n \"prompt\": \"A right circular cylinder is inscribed within a cube. The volume of the cube is 64 cubic centimeters. What is the volume of the cylinder, in cubic centimeters?\",\n \"choices\": {\n \"A\": \"$8\\\\pi$\",\n \"B\": \"$16\\\\pi$\",\n \"C\": \"$32\\\\pi$\",\n \"D\": \"$64\\\\pi$\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"B\"\n },\n {\n \"id\": \"sat99-math-m1-q17\",\n \"number\": 17,\n \"type\": \"geometry_trigonometry\",\n \"prompt\": \"A circle has a radius of 10. The area of a sector of the circle is $10\\\\pi$. What is the measure of the sector's central angle, in degrees?\",\n \"answer_format\": \"student_produced\",\n \"correct_answer\": \"36\"\n }\n]\n```" }, "status": "ok" }, { "attempts": 1, "index": 3, "output": { "costUsd": 0, "model": "gemini-2.5-pro", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 5 SAT Math questions (of which 1 are student-produced-response and the rest are MCQ).\nType: statistics_probability | Test: 99 | Unit: 1 | Start Q#: 18\n\nMANDATORY: Your JSON array MUST contain exactly 5 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nMCQ Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"statistics_probability\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\nSPR Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"statistics_probability\",\"prompt\":\"...\",\"answer_format\":\"student_produced\",\"correct_answer\":\"numeric_value\"}\nPlace SPR questions at the END of the array.\n\nFollow instructions for your type:\n\n[algebra] (MCQ + 1 SPR)\n- Topics: linear equations (1 var, 2 vars), systems of 2 linear equations, linear inequalities, absolute value\n- Mix word problems and abstract algebra. Use $...$ notation.\n- MCQ distractors: sign errors, arithmetic mistakes, setup errors, partial answers\n- SPR answers: clean integers or simple fractions\n\n[functions_analysis] (MCQ + 1 SPR)\n- Topics: function evaluation, domain/range, transformations (f(x)+k, f(x-h)), graphical analysis, composition f(g(x)), piecewise\n- At least one conceptual insight question (e.g., (p,q) and (q,p) on same graph)\n- Describe graphs in text. All computations yield clean results.\n\n[geometry_trigonometry] (MCQ + 1 SPR)\n- Topics: circle equations, Pythagorean theorem, special right triangles, coordinate geometry, trig ratios, volume/surface area, angle relationships\n- Optional \"visual\" field: {\"type\":\"coordinate_plane\",\"description\":\"...\"}\n- Radical answers OK (e.g., $5\\sqrt{13}$). Multi-step reasoning required.\n\n[statistics_probability] (MCQ + 1 SPR)\n- Topics: mean/median/mode, standard deviation (conceptual), probability (conditional from tables), data interpretation, scatter plots/line of best fit, sampling/bias\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Distractors: P(A|B) vs P(B|A) confusion, mean vs median, causal from correlational\n\n[advanced_math] (MCQ + 1 SPR)\n- Topics: quadratics (factoring, discriminant), polynomials (zeros, end behavior), exponentials (growth/decay), rational expressions, radical expressions, non-linear systems\n- Hardest questions on the test. Multi-step reasoning required. Elegant algebraic insight preferred.\n- All expressions must be factorable with integer coefficients. SPR answers are clean numbers.\n\nVERIFY EVERY ANSWER: Solve each problem step-by-step before finalizing correct_answer.\nOutput ONLY a valid JSON array of EXACTLY 5 objects. No markdown. Count them before outputting.", "text": "```json\n[\n {\n \"id\": \"sat99-math-m1-q18\",\n \"number\": 18,\n \"type\": \"statistics_probability\",\n \"prompt\": \"A scatterplot shows the relationship between the number of hours a student practices a musical instrument per week and the number of mistakes made during a performance. The line of best fit is given by the equation $y = -2.6x + 15$, where $y$ is the predicted number of mistakes and $x$ is the number of hours practiced. What is the best interpretation of the slope of the line of best fit?\",\n \"choices\": {\n \"A\": \"For each additional hour of practice, the predicted number of mistakes decreases by 2.6.\",\n \"B\": \"A student who does not practice is predicted to make 15 mistakes.\",\n \"C\": \"For every 2.6 additional mistakes made, the predicted number of practice hours decreases by 1.\",\n \"D\": \"The maximum possible number of mistakes is 15.\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"A\"\n },\n {\n \"id\": \"sat99-math-m1-q19\",\n \"number\": 19,\n \"type\": \"statistics_probability\",\n \"visual\": {\n \"type\": \"table\",\n \"title\": \"Participation in After-School Programs\",\n \"headers\": [\"\", \"Sports\", \"Arts\", \"Total\"],\n \"rows\": [\n [\"Juniors\", 85, 65, 150],\n [\"Seniors\", 95, 55, 150],\n [\"Total\", 180, 120, 300]\n ]\n },\n \"prompt\": \"The table above shows the distribution of 300 juniors and seniors at a high school based on their participation in after-school programs. If a student who participates in sports is selected at random, what is the probability that the student is a senior?\",\n \"choices\": {\n \"A\": \"$\\\\frac{95}{300}$\",\n \"B\": \"$\\\\frac{150}{300}$\",\n \"C\": \"$\\\\frac{95}{150}$\",\n \"D\": \"$\\\\frac{95}{180}$\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"D\"\n },\n {\n \"id\": \"sat99-math-m1-q20\",\n \"number\": 20,\n \"type\": \"statistics_probability\",\n \"prompt\": \"A data set consists of 9 integers. The numbers are 10, 12, 12, 13, 15, 17, 19, 20, and 45. Which of the following statements is true about the mean, median, and mode of this data set?\",\n \"choices\": {\n \"A\": \"The mean is greater than the median.\",\n \"B\": \"The median is greater than the mean.\",\n \"C\": \"The mode is greater than the mean.\",\n \"D\": \"The mean, median, and mode are all equal.\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"A\"\n },\n {\n \"id\": \"sat99-math-m1-q21\",\n \"number\": 21,\n \"type\": \"statistics_probability\",\n \"prompt\": \"An environmental scientist wants to estimate the proportion of trees in a large forest that are infected with a certain fungus. She divides the forest into 100 equally sized plots, randomly selects 10 plots, and examines every tree within those 10 plots. Which of the following best describes this sampling method?\",\n \"choices\": {\n \"A\": \"Simple random sampling\",\n \"B\": \"Cluster sampling\",\n \"C\": \"Stratified random sampling\",\n \"D\": \"Convenience sampling\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"B\"\n },\n {\n \"id\": \"sat99-math-m1-q22\",\n \"number\": 22,\n \"type\": \"statistics_probability\",\n \"visual\": {\n \"type\": \"table\",\n \"title\": \"Vehicle Types at a Service Center\",\n \"headers\": [\"\", \"Requires Oil Change\", \"Requires Tire Rotation\", \"Total\"],\n \"rows\": [\n [\"Sedan\", 50, 30, 80],\n [\"SUV\", 40, 35, 75],\n [\"Total\", 90, 65, 155]\n ]\n },\n \"prompt\": \"The table above shows the types of vehicles and the primary service requested at an auto service center on one day. Based on the table, what proportion of sedans required an oil change? (Express your answer as a decimal or fraction).\",\n \"answer_format\": \"student_produced\",\n \"correct_answer\": \"5/8\"\n }\n]\n```" }, "status": "ok" }, { "attempts": 2, "index": 4, "output": { "costUsd": 0, "model": "gemini-2.5-pro", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 5 SAT Math questions (of which 2 are student-produced-response and the rest are MCQ).\nType: advanced_math | Test: 99 | Unit: 1 | Start Q#: 23\n\nMANDATORY: Your JSON array MUST contain exactly 5 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nMCQ Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"advanced_math\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\nSPR Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"advanced_math\",\"prompt\":\"...\",\"answer_format\":\"student_produced\",\"correct_answer\":\"numeric_value\"}\nPlace SPR questions at the END of the array.\n\nFollow instructions for your type:\n\n[algebra] (MCQ + 2 SPR)\n- Topics: linear equations (1 var, 2 vars), systems of 2 linear equations, linear inequalities, absolute value\n- Mix word problems and abstract algebra. Use $...$ notation.\n- MCQ distractors: sign errors, arithmetic mistakes, setup errors, partial answers\n- SPR answers: clean integers or simple fractions\n\n[functions_analysis] (MCQ + 2 SPR)\n- Topics: function evaluation, domain/range, transformations (f(x)+k, f(x-h)), graphical analysis, composition f(g(x)), piecewise\n- At least one conceptual insight question (e.g., (p,q) and (q,p) on same graph)\n- Describe graphs in text. All computations yield clean results.\n\n[geometry_trigonometry] (MCQ + 2 SPR)\n- Topics: circle equations, Pythagorean theorem, special right triangles, coordinate geometry, trig ratios, volume/surface area, angle relationships\n- Optional \"visual\" field: {\"type\":\"coordinate_plane\",\"description\":\"...\"}\n- Radical answers OK (e.g., $5\\sqrt{13}$). Multi-step reasoning required.\n\n[statistics_probability] (MCQ + 2 SPR)\n- Topics: mean/median/mode, standard deviation (conceptual), probability (conditional from tables), data interpretation, scatter plots/line of best fit, sampling/bias\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Distractors: P(A|B) vs P(B|A) confusion, mean vs median, causal from correlational\n\n[advanced_math] (MCQ + 2 SPR)\n- Topics: quadratics (factoring, discriminant), polynomials (zeros, end behavior), exponentials (growth/decay), rational expressions, radical expressions, non-linear systems\n- Hardest questions on the test. Multi-step reasoning required. Elegant algebraic insight preferred.\n- All expressions must be factorable with integer coefficients. SPR answers are clean numbers.\n\nVERIFY EVERY ANSWER: Solve each problem step-by-step before finalizing correct_answer.\nOutput ONLY a valid JSON array of EXACTLY 5 objects. No markdown. Count them before outputting.", "text": "```json\n[\n {\n \"id\": \"sat99-math-m1-q23\",\n \"number\": 23,\n \"type\": \"advanced_math\",\n \"prompt\": \"The quadratic equation $kx^2 - 6x + 3 = 0$, where $k$ is a constant, has exactly one real solution. What is the value of $k$?\",\n \"choices\": {\n \"A\": \"-3\",\n \"B\": \"0\",\n \"C\": \"3\",\n \"D\": \"9\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"C\"\n },\n {\n \"id\": \"sat99-math-m1-q24\",\n \"number\": 24,\n \"type\": \"advanced_math\",\n \"prompt\": \"The polynomial $P(x) = x^3 - 2x^2 - 5x + 6$ has a zero at $x=1$. What is the sum of the other two zeros of the polynomial?\",\n \"choices\": {\n \"A\": \"-1\",\n \"B\": \"1\",\n \"C\": \"2\",\n \"D\": \"5\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"B\"\n },\n {\n \"id\": \"sat99-math-m1-q25\",\n \"number\": 25,\n \"type\": \"advanced_math\",\n \"prompt\": \"If $x > 2$, which of the following expressions is equivalent to $\\\\frac{x - \\\\frac{4}{x}}{1 + \\\\frac{2}{x}}$?\",\n \"choices\": {\n \"A\": \"x - 2\",\n \"B\": \"x + 2\",\n \"C\": \"\\\\frac{x-4}{3}\",\n \"D\": \"x - 4\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"A\"\n },\n {\n \"id\": \"sat99-math-m1-q26\",\n \"number\": 26,\n \"type\": \"advanced_math\",\n \"prompt\": \"A system of two equations is given by $y = x^2 - 5x + 7$ and $y - x = 2$. If $(x, y)$ is a solution to the system, what is one possible value for $x$?\",\n \"answer_format\": \"student_produced\",\n \"correct_answer\": \"5\"\n },\n {\n \"id\": \"sat99-math-m1-q27\",\n \"number\": 27,\n \"type\": \"advanced_math\",\n \"prompt\": \"If $n > 0$ and $\\\\sqrt[3]{n^2} \\\\cdot n^k = n^{5/3}$, what is the value of $k$?\",\n \"answer_format\": \"student_produced\",\n \"correct_answer\": \"1\"\n }\n]\n```" }, "status": "ok" }, { "attempts": 1, "index": 5, "output": { "costUsd": 0, "model": "gemini-2.5-pro", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 7 SAT Math questions (of which 2 are student-produced-response and the rest are MCQ).\nType: algebra | Test: 99 | Unit: 2 | Start Q#: 1\n\nMANDATORY: Your JSON array MUST contain exactly 7 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nMCQ Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"algebra\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\nSPR Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"algebra\",\"prompt\":\"...\",\"answer_format\":\"student_produced\",\"correct_answer\":\"numeric_value\"}\nPlace SPR questions at the END of the array.\n\nFollow instructions for your type:\n\n[algebra] (MCQ + 2 SPR)\n- Topics: linear equations (1 var, 2 vars), systems of 2 linear equations, linear inequalities, absolute value\n- Mix word problems and abstract algebra. Use $...$ notation.\n- MCQ distractors: sign errors, arithmetic mistakes, setup errors, partial answers\n- SPR answers: clean integers or simple fractions\n\n[functions_analysis] (MCQ + 2 SPR)\n- Topics: function evaluation, domain/range, transformations (f(x)+k, f(x-h)), graphical analysis, composition f(g(x)), piecewise\n- At least one conceptual insight question (e.g., (p,q) and (q,p) on same graph)\n- Describe graphs in text. All computations yield clean results.\n\n[geometry_trigonometry] (MCQ + 2 SPR)\n- Topics: circle equations, Pythagorean theorem, special right triangles, coordinate geometry, trig ratios, volume/surface area, angle relationships\n- Optional \"visual\" field: {\"type\":\"coordinate_plane\",\"description\":\"...\"}\n- Radical answers OK (e.g., $5\\sqrt{13}$). Multi-step reasoning required.\n\n[statistics_probability] (MCQ + 2 SPR)\n- Topics: mean/median/mode, standard deviation (conceptual), probability (conditional from tables), data interpretation, scatter plots/line of best fit, sampling/bias\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Distractors: P(A|B) vs P(B|A) confusion, mean vs median, causal from correlational\n\n[advanced_math] (MCQ + 2 SPR)\n- Topics: quadratics (factoring, discriminant), polynomials (zeros, end behavior), exponentials (growth/decay), rational expressions, radical expressions, non-linear systems\n- Hardest questions on the test. Multi-step reasoning required. Elegant algebraic insight preferred.\n- All expressions must be factorable with integer coefficients. SPR answers are clean numbers.\n\nVERIFY EVERY ANSWER: Solve each problem step-by-step before finalizing correct_answer.\nOutput ONLY a valid JSON array of EXACTLY 7 objects. No markdown. Count them before outputting.", "text": "```json\n[\n {\n \"id\": \"sat99-math-m2-q1\",\n \"number\": 1,\n \"type\": \"algebra\",\n \"prompt\": \"A car rental agency charges a flat daily fee of $40 plus an additional charge of $0.15 per mile driven. If a customer's total bill for a one-day rental was $98.50, how many miles did the customer drive?\",\n \"choices\": {\n \"A\": \"267\",\n \"B\": \"39\",\n \"C\": \"390\",\n \"D\": \"657\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"C\"\n },\n {\n \"id\": \"sat99-math-m2-q2\",\n \"number\": 2,\n \"type\": \"algebra\",\n \"prompt\": \"If $|2x - 3| = 11$, what is one possible value of $x$?\",\n \"choices\": {\n \"A\": \"7\",\n \"B\": \"4\",\n \"C\": \"-7\",\n \"D\": \"-4\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"D\"\n },\n {\n \"id\": \"sat99-math-m2-q3\",\n \"number\": 3,\n \"type\": \"algebra\",\n \"prompt\": \"At a food truck, 2 tacos and 1 burrito cost a total of $11.00. At the same food truck, 1 taco and 3 burritos cost a total of $18.00. What is the cost of one burrito?\",\n \"choices\": {\n \"A\": \"$2.50\",\n \"B\": \"$3.00\",\n \"C\": \"$5.50\",\n \"D\": \"$6.00\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"C\"\n },\n {\n \"id\": \"sat99-math-m2-q4\",\n \"number\": 4,\n \"type\": \"algebra\",\n \"prompt\": \"Which of the following values of $x$ is a solution to the inequality $8 - 3x < -7$?\",\n \"choices\": {\n \"A\": \"6\",\n \"B\": \"5\",\n \"C\": \"4\",\n \"D\": \"-5\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"A\"\n },\n {\n \"id\": \"sat99-math-m2-q5\",\n \"number\": 5,\n \"type\": \"algebra\",\n \"prompt\": \"A small business's weekly profit, $P$, in dollars, is modeled by the function $P(x) = 150x - 2700$, where $x$ represents the number of items sold. What is the minimum number of items the business must sell in a week to realize a profit?\",\n \"choices\": {\n \"A\": \"19\",\n \"B\": \"18\",\n \"C\": \"17\",\n \"D\": \"16\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"A\"\n },\n {\n \"id\": \"sat99-math-m2-q6\",\n \"number\": 6,\n \"type\": \"algebra\",\n \"prompt\": \"If $3x + 2y = 19$ and $x - y = 3$, what is the value of $x$?\",\n \"answer_format\": \"student_produced\",\n \"correct_answer\": \"5\"\n },\n {\n \"id\": \"sat99-math-m2-q7\",\n \"number\": 7,\n \"type\": \"algebra\",\n \"prompt\": \"If $\\\\frac{2}{3}(x - 4) = \\\\frac{1}{2}(x + 2)$, what is the value of $x$?\",\n \"answer_format\": \"student_produced\",\n \"correct_answer\": \"22\"\n }\n]\n```" }, "status": "ok" }, { "attempts": 1, "index": 6, "output": { "costUsd": 0, "model": "gemini-2.5-pro", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 5 SAT Math questions (of which 1 are student-produced-response and the rest are MCQ).\nType: functions_analysis | Test: 99 | Unit: 2 | Start Q#: 8\n\nMANDATORY: Your JSON array MUST contain exactly 5 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nMCQ Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"functions_analysis\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\nSPR Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"functions_analysis\",\"prompt\":\"...\",\"answer_format\":\"student_produced\",\"correct_answer\":\"numeric_value\"}\nPlace SPR questions at the END of the array.\n\nFollow instructions for your type:\n\n[algebra] (MCQ + 1 SPR)\n- Topics: linear equations (1 var, 2 vars), systems of 2 linear equations, linear inequalities, absolute value\n- Mix word problems and abstract algebra. Use $...$ notation.\n- MCQ distractors: sign errors, arithmetic mistakes, setup errors, partial answers\n- SPR answers: clean integers or simple fractions\n\n[functions_analysis] (MCQ + 1 SPR)\n- Topics: function evaluation, domain/range, transformations (f(x)+k, f(x-h)), graphical analysis, composition f(g(x)), piecewise\n- At least one conceptual insight question (e.g., (p,q) and (q,p) on same graph)\n- Describe graphs in text. All computations yield clean results.\n\n[geometry_trigonometry] (MCQ + 1 SPR)\n- Topics: circle equations, Pythagorean theorem, special right triangles, coordinate geometry, trig ratios, volume/surface area, angle relationships\n- Optional \"visual\" field: {\"type\":\"coordinate_plane\",\"description\":\"...\"}\n- Radical answers OK (e.g., $5\\sqrt{13}$). Multi-step reasoning required.\n\n[statistics_probability] (MCQ + 1 SPR)\n- Topics: mean/median/mode, standard deviation (conceptual), probability (conditional from tables), data interpretation, scatter plots/line of best fit, sampling/bias\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Distractors: P(A|B) vs P(B|A) confusion, mean vs median, causal from correlational\n\n[advanced_math] (MCQ + 1 SPR)\n- Topics: quadratics (factoring, discriminant), polynomials (zeros, end behavior), exponentials (growth/decay), rational expressions, radical expressions, non-linear systems\n- Hardest questions on the test. Multi-step reasoning required. Elegant algebraic insight preferred.\n- All expressions must be factorable with integer coefficients. SPR answers are clean numbers.\n\nVERIFY EVERY ANSWER: Solve each problem step-by-step before finalizing correct_answer.\nOutput ONLY a valid JSON array of EXACTLY 5 objects. No markdown. Count them before outputting.", "text": "```json\n[\n {\n \"id\": \"sat99-math-m2-q8\",\n \"number\": 8,\n \"type\": \"functions_analysis\",\n \"prompt\": \"For the function $f$ defined by $f(x) = 2x^2 - 5x + 1$, what is the value of $f(3) - f(-1)$?\",\n \"choices\": {\n \"A\": \"-4\",\n \"B\": \"-1\",\n \"C\": \"4\",\n \"D\": \"12\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"A\"\n },\n {\n \"id\": \"sat99-math-m2-q9\",\n \"number\": 9,\n \"type\": \"functions_analysis\",\n \"prompt\": \"The graph of a function $g$ is shown in the $xy$-plane. The function $h$ is defined by $h(x) = g(x-2) + 3$. If the point $(-1, 4)$ is on the graph of $g$, what is a point on the graph of $h$?\",\n \"choices\": {\n \"A\": \"(-3, 7)\",\n \"B\": \"(-3, 1)\",\n \"C\": \"(1, 7)\",\n \"D\": \"(1, 1)\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"C\"\n },\n {\n \"id\": \"sat99-math-m2-q10\",\n \"number\": 10,\n \"type\": \"functions_analysis\",\n \"prompt\": \"The function $f$ is an even function, meaning $f(x) = f(-x)$ for all values of $x$. The function $g$ is an odd function, meaning $g(-x) = -g(x)$ for all values of $x$. If $f(2) = 5$ and $g(-3) = -4$, what is the value of $f(-2) + g(3)$?\",\n \"choices\": {\n \"A\": \"-9\",\n \"B\": \"1\",\n \"C\": \"9\",\n \"D\": \"The value cannot be determined from the information given.\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"C\"\n },\n {\n \"id\": \"sat99-math-m2-q11\",\n \"number\": 11,\n \"type\": \"functions_analysis\",\n \"prompt\": \"Let the functions $p$ and $q$ be defined by $p(x) = x^2 - 4$ and $q(x) = 2x + 1$. For what positive value of $x$ is $p(q(x)) = 0$?\",\n \"choices\": {\n \"A\": \"$\\\\frac{1}{2}$\",\n \"B\": \"$\\\\frac{3}{2}$\",\n \"C\": \"$2$\",\n \"D\": \"$5$\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"A\"\n },\n {\n \"id\": \"sat99-math-m2-q12\",\n \"number\": 12,\n \"type\": \"functions_analysis\",\n \"prompt\": \"The function $f$ is defined below.\\n$f(x) = \\\\begin{cases} x^2+3 & \\\\text{if } x < 1 \\\\\\\\ 2x-1 & \\\\text{if } x \\\\ge 1 \\\\end{cases}$\\nWhat is the value of $f(f(0))$?\",\n \"answer_format\": \"student_produced\",\n \"correct_answer\": \"5\"\n }\n]\n```" }, "status": "ok" }, { "attempts": 1, "index": 7, "output": { "costUsd": 0, "model": "gemini-2.5-pro", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 5 SAT Math questions (of which 1 are student-produced-response and the rest are MCQ).\nType: geometry_trigonometry | Test: 99 | Unit: 2 | Start Q#: 13\n\nMANDATORY: Your JSON array MUST contain exactly 5 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nMCQ Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"geometry_trigonometry\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\nSPR Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"geometry_trigonometry\",\"prompt\":\"...\",\"answer_format\":\"student_produced\",\"correct_answer\":\"numeric_value\"}\nPlace SPR questions at the END of the array.\n\nFollow instructions for your type:\n\n[algebra] (MCQ + 1 SPR)\n- Topics: linear equations (1 var, 2 vars), systems of 2 linear equations, linear inequalities, absolute value\n- Mix word problems and abstract algebra. Use $...$ notation.\n- MCQ distractors: sign errors, arithmetic mistakes, setup errors, partial answers\n- SPR answers: clean integers or simple fractions\n\n[functions_analysis] (MCQ + 1 SPR)\n- Topics: function evaluation, domain/range, transformations (f(x)+k, f(x-h)), graphical analysis, composition f(g(x)), piecewise\n- At least one conceptual insight question (e.g., (p,q) and (q,p) on same graph)\n- Describe graphs in text. All computations yield clean results.\n\n[geometry_trigonometry] (MCQ + 1 SPR)\n- Topics: circle equations, Pythagorean theorem, special right triangles, coordinate geometry, trig ratios, volume/surface area, angle relationships\n- Optional \"visual\" field: {\"type\":\"coordinate_plane\",\"description\":\"...\"}\n- Radical answers OK (e.g., $5\\sqrt{13}$). Multi-step reasoning required.\n\n[statistics_probability] (MCQ + 1 SPR)\n- Topics: mean/median/mode, standard deviation (conceptual), probability (conditional from tables), data interpretation, scatter plots/line of best fit, sampling/bias\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Distractors: P(A|B) vs P(B|A) confusion, mean vs median, causal from correlational\n\n[advanced_math] (MCQ + 1 SPR)\n- Topics: quadratics (factoring, discriminant), polynomials (zeros, end behavior), exponentials (growth/decay), rational expressions, radical expressions, non-linear systems\n- Hardest questions on the test. Multi-step reasoning required. Elegant algebraic insight preferred.\n- All expressions must be factorable with integer coefficients. SPR answers are clean numbers.\n\nVERIFY EVERY ANSWER: Solve each problem step-by-step before finalizing correct_answer.\nOutput ONLY a valid JSON array of EXACTLY 5 objects. No markdown. Count them before outputting.", "text": "```json\n[\n {\n \"id\": \"sat99-math-m2-q13\",\n \"number\": 13,\n \"type\": \"geometry_trigonometry\",\n \"prompt\": \"A circle in the $xy$-plane has its center at $(3, -5)$. The point $(8, 7)$ lies on the circle. Which of the following is the equation of the circle?\",\n \"choices\": {\n \"A\": \"$(x-3)^2 + (y+5)^2 = 13$\",\n \"B\": \"$(x-8)^2 + (y-7)^2 = 169$\",\n \"C\": \"$(x+3)^2 + (y-5)^2 = 169$\",\n \"D\": \"$(x-3)^2 + (y+5)^2 = 169$\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"D\"\n },\n {\n \"id\": \"sat99-math-m2-q14\",\n \"number\": 14,\n \"type\": \"geometry_trigonometry\",\n \"prompt\": \"In right triangle $ABC$, the measure of angle $C$ is $90^{\\\\circ}$. If $\\\\sin(A) = \\\\frac{8}{17}$, what is the value of $\\\\cos(B)$?\",\n \"choices\": {\n \"A\": \"$\\\\frac{15}{17}$\",\n \"B\": \"$\\\\frac{8}{15}$\",\n \"C\": \"$\\\\frac{8}{17}$\",\n \"D\": \"$\\\\frac{17}{8}$\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"C\"\n },\n {\n \"id\": \"sat99-math-m2-q15\",\n \"number\": 15,\n \"type\": \"geometry_trigonometry\",\n \"prompt\": \"A cylindrical grain silo has a radius of 10 feet and a height of 40 feet. The top of the silo is a hemisphere with the same radius. What is the total volume of the silo in cubic feet?\",\n \"choices\": {\n \"A\": \"$4000\\\\pi$\",\n \"B\": \"$\\\\frac{14000}{3}\\\\pi$\",\n \"C\": \"$\\\\frac{16000}{3}\\\\pi$\",\n \"D\": \"$4400\\\\pi$\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"B\"\n },\n {\n \"id\": \"sat99-math-m2-q16\",\n \"number\": 16,\n \"type\": \"geometry_trigonometry\",\n \"prompt\": \"In the $xy$-plane, line $L$ is defined by the equation $y = -\\\\frac{1}{3}x + 5$. Line $M$ is perpendicular to line $L$ and passes through the point $(-2, 2)$. What is the $y$-intercept of line $M$?\",\n \"choices\": {\n \"A\": \"8\",\n \"B\": \"-8\",\n \"C\": \"3\",\n \"D\": \"$-\\\\frac{1}{3}$\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"A\"\n },\n {\n \"id\": \"sat99-math-m2-q17\",\n \"number\": 17,\n \"type\": \"geometry_trigonometry\",\n \"prompt\": \"A ladder is placed against a vertical wall. The base of the ladder is 7 feet away from the base of the wall, and the ladder makes an angle of $60^{\\\\circ}$ with the ground. What is the length of the ladder in feet?\",\n \"answer_format\": \"student_produced\",\n \"correct_answer\": \"14\"\n }\n]\n```" }, "status": "ok" }, { "attempts": 1, "index": 8, "output": { "costUsd": 0, "model": "gemini-2.5-pro", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 5 SAT Math questions (of which 1 are student-produced-response and the rest are MCQ).\nType: statistics_probability | Test: 99 | Unit: 2 | Start Q#: 18\n\nMANDATORY: Your JSON array MUST contain exactly 5 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nMCQ Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"statistics_probability\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\nSPR Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"statistics_probability\",\"prompt\":\"...\",\"answer_format\":\"student_produced\",\"correct_answer\":\"numeric_value\"}\nPlace SPR questions at the END of the array.\n\nFollow instructions for your type:\n\n[algebra] (MCQ + 1 SPR)\n- Topics: linear equations (1 var, 2 vars), systems of 2 linear equations, linear inequalities, absolute value\n- Mix word problems and abstract algebra. Use $...$ notation.\n- MCQ distractors: sign errors, arithmetic mistakes, setup errors, partial answers\n- SPR answers: clean integers or simple fractions\n\n[functions_analysis] (MCQ + 1 SPR)\n- Topics: function evaluation, domain/range, transformations (f(x)+k, f(x-h)), graphical analysis, composition f(g(x)), piecewise\n- At least one conceptual insight question (e.g., (p,q) and (q,p) on same graph)\n- Describe graphs in text. All computations yield clean results.\n\n[geometry_trigonometry] (MCQ + 1 SPR)\n- Topics: circle equations, Pythagorean theorem, special right triangles, coordinate geometry, trig ratios, volume/surface area, angle relationships\n- Optional \"visual\" field: {\"type\":\"coordinate_plane\",\"description\":\"...\"}\n- Radical answers OK (e.g., $5\\sqrt{13}$). Multi-step reasoning required.\n\n[statistics_probability] (MCQ + 1 SPR)\n- Topics: mean/median/mode, standard deviation (conceptual), probability (conditional from tables), data interpretation, scatter plots/line of best fit, sampling/bias\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Distractors: P(A|B) vs P(B|A) confusion, mean vs median, causal from correlational\n\n[advanced_math] (MCQ + 1 SPR)\n- Topics: quadratics (factoring, discriminant), polynomials (zeros, end behavior), exponentials (growth/decay), rational expressions, radical expressions, non-linear systems\n- Hardest questions on the test. Multi-step reasoning required. Elegant algebraic insight preferred.\n- All expressions must be factorable with integer coefficients. SPR answers are clean numbers.\n\nVERIFY EVERY ANSWER: Solve each problem step-by-step before finalizing correct_answer.\nOutput ONLY a valid JSON array of EXACTLY 5 objects. No markdown. Count them before outputting.", "text": "```json\n[\n {\n \"id\": \"sat99-math-m2-q18\",\n \"number\": 18,\n \"type\": \"statistics_probability\",\n \"prompt\": \"A market research firm surveyed 200 owners of either a sedan or an SUV. Each owner was asked whether their next vehicle purchase would be a new or used vehicle. The results are partially summarized in the table below. Of those surveyed, 60% owned a sedan.\",\n \"visual\": {\n \"type\": \"table\",\n \"title\": \"Next Vehicle Purchase Plan\",\n \"headers\": [\"Current Vehicle\", \"New\", \"Used\", \"Total\"],\n \"rows\": [\n [\"Sedan\", 72, \"\", \"\"],\n [\"SUV\", \"\", \"\", \"\"],\n [\"Total\", 110, 90, 200]\n ]\n },\n \"prompt_continue\": \"Based on the table, what is the probability that a person who currently owns an SUV will plan to purchase a new vehicle?\",\n \"choices\": {\n \"A\": \"0.34\",\n \"B\": \"0.475\",\n \"C\": \"0.525\",\n \"D\": \"0.65\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"B\"\n },\n {\n \"id\": \"sat99-math-m2-q19\",\n \"number\": 19,\n \"type\": \"statistics_probability\",\n \"prompt\": \"The dot plot above shows the number of goals scored by a soccer team in 15 games. If the game with the highest number of goals is removed from the data, which of the following statements about the mean and median of the resulting data is true?\",\n \"visual\": {\n \"type\": \"dot_plot\",\n \"description\": \"A dot plot titled 'Goals Scored in 15 Games'. The x-axis is labeled 'Number of Goals'. There are 3 dots at 0, 5 dots at 1, 4 dots at 2, 2 dots at 3, and 1 dot at 6.\"\n },\n \"choices\": {\n \"A\": \"The mean will decrease, and the median will remain the same.\",\n \"B\": \"The mean will decrease, and the median will decrease.\",\n \"C\": \"The mean will remain the same, and the median will decrease.\",\n \"D\": \"Both the mean and the median will remain the same.\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"A\"\n },\n {\n \"id\": \"sat99-math-m2-q20\",\n \"number\": 20,\n \"type\": \"statistics_probability\",\n \"prompt\": \"A sample of 50 trees was selected from a large forest. The height, in feet, and the trunk diameter, in inches, was measured for each tree. The data are shown in the scatterplot, along with a line of best fit. The equation of the line is $h = 3.8d + 15.2$, where $h$ is the height and $d$ is the trunk diameter.\",\n \"visual\": {\n \"type\": \"scatter_plot\",\n \"description\": \"A scatterplot titled 'Tree Height vs. Trunk Diameter' showing a positive linear association. The x-axis is 'Trunk Diameter (inches)' from 5 to 25. The y-axis is 'Height (feet)' from 20 to 120. A line of best fit with a positive slope is drawn through the data points.\"\n },\n \"prompt_continue\": \"Which of the following is the best interpretation of the slope of the line of best fit?\",\n \"choices\": {\n \"A\": \"For a tree with a trunk diameter of 0 inches, the predicted height is 3.8 feet.\",\n \"B\": \"For every 1-foot increase in a tree's height, the predicted trunk diameter increases by 3.8 inches.\",\n \"C\": \"For every 3.8-inch increase in a tree's trunk diameter, the predicted height increases by 1 foot.\",\n \"D\": \"For every 1-inch increase in a tree's trunk diameter, the predicted height increases by 3.8 feet.\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"D\"\n },\n {\n \"id\": \"sat99-math-m2-q21\",\n \"number\": 21,\n \"type\": \"statistics_probability\",\n \"prompt\": \"The test scores for two history classes are summarized by the box plots below.\",\n \"visual\": {\n \"type\": \"box_plot\",\n \"description\": \"Two box plots on a single axis from 50 to 100. 'Class A': min=55, Q1=65, median=75, Q3=80, max=90. 'Class B': min=60, Q1=75, median=80, Q3=85, max=90. The box for Class A is wider than the box for Class B, indicating a larger interquartile range.\"\n },\n \"prompt_continue\": \"Based on the box plots, which of the following must be true?\",\n \"choices\": {\n \"A\": \"The mean score of Class A is greater than the mean score of Class B.\",\n \"B\": \"The range of scores for Class A is greater than the range of scores for Class B.\",\n \"C\": \"The median score of Class B is greater than the median score of Class A.\",\n \"D\": \"There are more students in Class B than in Class A.\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"C\"\n },\n {\n \"id\": \"sat99-math-m2-q22\",\n \"number\": 22,\n \"type\": \"statistics_probability\",\n \"prompt\": \"A dataset consists of 7 distinct integers. The median of the dataset is 12, the minimum value is 5, and the range is 14. What is the maximum possible value of the mean of the dataset, rounded to the nearest tenth?\",\n \"answer_format\": \"student_produced\",\n \"correct_answer\": \"14.3\"\n }\n]\n```" }, "status": "ok" }, { "attempts": 1, "index": 9, "output": { "costUsd": 0, "model": "gemini-2.5-pro", "prompt": "You are one of the world's foremost standardized test question writers with 25 years of College Board SAT experience. Your questions pass psychometric review on first submission.\n\nABSOLUTE RULES (violation = rejection):\n1. Output ONLY a valid JSON array. No markdown fences. No commentary before/after.\n2. The array MUST contain EXACTLY the requested count of questions — not fewer, not more.\n3. Every MCQ has EXACTLY 4 choices keyed \"A\", \"B\", \"C\", \"D\". NEVER add an \"E\" choice. NEVER leave any choice empty.\n4. Each question has EXACTLY ONE defensible correct answer.\n5. Every question is 100% original (no recycled phrasing from known tests).\n6. Every distractor represents a real, common student mistake — not nonsense or trivially wrong.\n7. Zero cultural, racial, gender, or socioeconomic bias.\n8. Use $...$ LaTeX for ALL math notation (variables, equations, fractions, exponents).\n9. Reading passages must feel like genuine published academic/literary excerpts.\n10. VERIFY your own answer key before finalizing — mentally solve/evaluate every question.\n\nDISTRACTOR QUALITY (critical):\n- Vocabulary: near-synonyms from the same semantic field that fail in THIS specific context\n- Reading: plausible misreadings (wrong scope, surface-level, opposite, over-generalization)\n- Math MCQ: each wrong answer = a specific common error (sign flip, wrong formula, off-by-one, partial simplification)\n- Grammar: each wrong choice introduces ONE specific grammatical error\n- At least 2 distractors should require careful analysis to eliminate.\n\nANSWER DISTRIBUTION (critical): Distribute correct answers roughly evenly across A/B/C/D.\n- In a batch of 4+ questions, no letter should be correct more than (count/4)+1 times.\n- Never have 3+ consecutive same-letter answers.\n- Before finalizing, count your answer letters; if skewed (e.g., 5 Bs in a batch of 8), rebalance.\n\nCOUNT CHECK (critical): Before outputting, verify your array has the EXACT requested count. Missing or extra items = rejected.\n\nGenerate EXACTLY 5 SAT Math questions (of which 2 are student-produced-response and the rest are MCQ).\nType: advanced_math | Test: 99 | Unit: 2 | Start Q#: 23\n\nMANDATORY: Your JSON array MUST contain exactly 5 question objects.\nEach MCQ has exactly 4 choices: A, B, C, D (no E, no empty values).\n\nMCQ Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"advanced_math\",\"prompt\":\"...\",\"choices\":{\"A\":\"...\",\"B\":\"...\",\"C\":\"...\",\"D\":\"...\"},\"answer_format\":\"mcq\",\"correct_answer\":\"X\"}\nSPR Schema: {\"id\":\"sat{testNumber}-math-m{mod}-q{N}\",\"number\":N,\"type\":\"advanced_math\",\"prompt\":\"...\",\"answer_format\":\"student_produced\",\"correct_answer\":\"numeric_value\"}\nPlace SPR questions at the END of the array.\n\nFollow instructions for your type:\n\n[algebra] (MCQ + 2 SPR)\n- Topics: linear equations (1 var, 2 vars), systems of 2 linear equations, linear inequalities, absolute value\n- Mix word problems and abstract algebra. Use $...$ notation.\n- MCQ distractors: sign errors, arithmetic mistakes, setup errors, partial answers\n- SPR answers: clean integers or simple fractions\n\n[functions_analysis] (MCQ + 2 SPR)\n- Topics: function evaluation, domain/range, transformations (f(x)+k, f(x-h)), graphical analysis, composition f(g(x)), piecewise\n- At least one conceptual insight question (e.g., (p,q) and (q,p) on same graph)\n- Describe graphs in text. All computations yield clean results.\n\n[geometry_trigonometry] (MCQ + 2 SPR)\n- Topics: circle equations, Pythagorean theorem, special right triangles, coordinate geometry, trig ratios, volume/surface area, angle relationships\n- Optional \"visual\" field: {\"type\":\"coordinate_plane\",\"description\":\"...\"}\n- Radical answers OK (e.g., $5\\sqrt{13}$). Multi-step reasoning required.\n\n[statistics_probability] (MCQ + 2 SPR)\n- Topics: mean/median/mode, standard deviation (conceptual), probability (conditional from tables), data interpretation, scatter plots/line of best fit, sampling/bias\n- Include \"visual\" field (table) for at least half: {\"type\":\"table\",\"title\":\"...\",\"headers\":[...],\"rows\":[[...],...]}\n- Distractors: P(A|B) vs P(B|A) confusion, mean vs median, causal from correlational\n\n[advanced_math] (MCQ + 2 SPR)\n- Topics: quadratics (factoring, discriminant), polynomials (zeros, end behavior), exponentials (growth/decay), rational expressions, radical expressions, non-linear systems\n- Hardest questions on the test. Multi-step reasoning required. Elegant algebraic insight preferred.\n- All expressions must be factorable with integer coefficients. SPR answers are clean numbers.\n\nVERIFY EVERY ANSWER: Solve each problem step-by-step before finalizing correct_answer.\nOutput ONLY a valid JSON array of EXACTLY 5 objects. No markdown. Count them before outputting.", "text": "```json\n[\n {\n \"id\": \"sat99-math-m2-q23\",\n \"number\": 23,\n \"type\": \"advanced_math\",\n \"prompt\": \"The system of equations $y = x^2 - 4x + 7$ and $y = 2x + 2$ is graphed in the $xy$-plane. What is the sum of the $y$-coordinates of the two intersection points?\",\n \"choices\": {\n \"A\": \"6\",\n \"B\": \"12\",\n \"C\": \"18\",\n \"D\": \"24\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"C\"\n },\n {\n \"id\": \"sat99-math-m2-q24\",\n \"number\": 24,\n \"type\": \"advanced_math\",\n \"prompt\": \"A polynomial function $p(x)$ has zeros at $x=-1$ and $x=4$, where the zero at $x=4$ has a multiplicity of 2. The graph of $y=p(x)$ has a positive $y$-intercept. Which of the following could be the equation for $p(x)$?\",\n \"choices\": {\n \"A\": \"p(x) = (x+1)(x-4)^2\",\n \"B\": \"p(x) = (x-1)(x+4)^2\",\n \"C\": \"p(x) = -(x-1)(x+4)^2\",\n \"D\": \"p(x) = -(x+1)(x-4)^2\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"D\"\n },\n {\n \"id\": \"sat99-math-m2-q25\",\n \"number\": 25,\n \"type\": \"advanced_math\",\n \"prompt\": \"The function $f$ is defined by $f(x) = \\\\frac{2x^2 + 7x - 15}{x+5}$. The graph of $y=f(x)$ is a line with a point of discontinuity. What is the $y$-coordinate of the point of discontinuity?\",\n \"choices\": {\n \"A\": \"-13\",\n \"B\": \"-5\",\n \"C\": \"2\",\n \"D\": \"3\"\n },\n \"answer_format\": \"mcq\",\n \"correct_answer\": \"A\"\n },\n {\n \"id\": \"sat99-math-m2-q26\",\n \"number\": 26,\n \"type\": \"advanced_math\",\n \"prompt\": \"If $x > 0$, what value of $x$ is the solution to the equation $\\\\sqrt{2x+3} = x$?\",\n \"answer_format\": \"student_produced\",\n \"correct_answer\": \"3\"\n },\n {\n \"id\": \"sat99-math-m2-q27\",\n \"number\": 27,\n \"type\": \"advanced_math\",\n \"prompt\": \"The value of an investment is modeled by the function $V(t) = 100(1.5)^{\\\\frac{t}{4}}$, where $V(t)$ is the value of the investment in thousands of dollars and $t$ is the number of years since the initial investment was made. After how many years will the value of the investment be $225,000?\",\n \"answer_format\": \"student_produced\",\n \"correct_answer\": \"8\"\n }\n]\n```" }, "status": "ok" } ], "succeeded": 10 } } }

Noeud	Statut	Duree	Tentatives
Assemble Exam	Success	425ms	1
Build Batches	Success	255ms	1
Generate Math (Gemini 3 Pro)	Success	10m 30s	1
Generate RW (GPT 5.4)	Success	2m 49s	1
Test Number	Success	228ms	1
Output	Success	497ms	1

Run run_20260416Success