🏆 Quality winner: openrouter:deepseek/deepseek-v4-pro
The DeepSeek entry combines the most distinctive, mechanism-matching analogy (Soviet Tu-4 reverse-engineered from the B-29, which precisely mirrors how distilled models approximate but don't match originals) with a complete, fully answerable 10-question SAT-style quiz, a precise six-term glossary, and an original, engaging, non-truncated rewritten passage. It edges out c2b157 and e2507b, which are equally faithful and well-structured but have slightly more generic analogies, and beats the strong Claude/GPT entries because those passages are truncated. Weaker candidates either supplied only 1-3 quiz questions (f95e70, a0132a) or relied on generic chess/homework analogies (fa777b).
💰 Best value: openrouter:deepseek/deepseek-v4-pro
— quality 9/10 at 0.62¢ → value 8.7/10
Value = 0.7·quality + 0.3·cheapness. Quality stays dominant, so a cheap-but-weak output can't win on price alone.
| # | Model | Facts | Analogy | Age-fit | Clarity | Quiz | Glossary | Passage | Overall | Cost | Value | Notes |
|---|
| 🥇 |
openrouter:deepseek/deepseek-v4-pro
view ↗ |
9 | 9 | 9 | 9 | 9 | 9 | 9 |
9 |
0.62¢ |
8.7 |
Excellent fact fidelity with strong Cold War Tu-4/B-29 analogy sharing the reverse-engineering mechanism. Full 10-question quiz is fair and answerable from the rewritten passage. Glossary precise; passage is engaging and accurate. |
| 🥈 |
openrouter:minimax/minimax-m2.7
view ↗ |
9 | 8 | 9 | 9 | 9 | 9 | 9 |
9 |
0.62¢ |
8.7 |
Very faithful, well-organized with a sharp university/answer-key analogy and F1 comparison. Quiz answerable from the passage with good distractors and clean evidence-pairing. Glossary strong including AI safety guardrails. |
| 🥉 |
openrouter:moonshotai/kimi-k2.6
view ↗ |
9 | 8 | 9 | 9 | 9 | 8 | 9 |
9 |
9.62¢ |
6.6 |
Highly faithful, incorporates McGuire and the entity-list detail into an original, sophisticated passage. Conservatory and pharma analogies share the core mechanism well. Strong, trap-aware quiz with accurate evidence pairing. |
| 4 |
anthropic:claude-haiku-4-5-20251001
view ↗ |
9 | 8 | 9 | 9 | 9 | 9 | 8 |
9 |
2.57¢ |
7.5 |
Very faithful with a rich, well-structured passage (though truncated at the very end). Strong nuclear-arms and doping-scandal analogies capturing the mechanism. Excellent, trap-aware quiz and thorough glossary. |
| 5 |
openrouter:mistralai/mistral-large-2512
view ↗ |
8 | 8 | 9 | 9 | 8 | 8 | 8 |
8 |
0.85¢ |
8.0 |
Strong analogies (marathon sweatband, supply chain heist) and solid structure. Q5 asks about 'surreptitious' which appears in source but not clearly in the rewritten passage, a minor calibration issue. Otherwise faithful and readable. |
| 6 |
openrouter:google/gemini-3.1-flash-lite
view ↗ |
8 | 8 | 8 | 8 | 8 | 8 | 8 |
8 |
0.59¢ |
8.0 |
Solid chef-recipe and blueprint analogies that fit the mechanism. Faithful passage and well-constructed 10-question quiz with reasonable evidence pairing. Clear, scannable structure. |
| 7 |
openrouter:z-ai/glm-5-turbo
view ↗ |
8 | 8 | 9 | 8 | 8 | 9 | 7 |
8 |
3.40¢ |
6.8 |
Strong analogies and detailed trap-labeled quiz, but includes a fabricated vocab question ('sharp' in 'drew a sharp distinction') that relies on the rewritten passage's own added phrasing, and the passage is truncated. Otherwise rich and well-grounded. |
| 8 |
openrouter:x-ai/grok-4.3
view ↗ |
9 | 7 | 8 | 8 | 8 | 8 | 8 |
8 |
1.64¢ |
7.4 |
Faithful and complete with a clean, original passage incorporating McGuire. Analogies (playbook cameras, reverse-engineering) are adequate but somewhat generic. Solid trap-aware quiz answerable from passage. |
| 9 |
openrouter:bytedance-seed/seed-2.0-lite
view ↗ |
8 | 7 | 8 | 8 | 8 | 8 | 8 |
8 |
1.73¢ |
7.4 |
Faithful, complete passage and strong SAT-style quiz with good traps. MIT-lab and gaming analogies are decent but a bit forced. Slightly overstates figures ('$1 trillion industry') but broadly accurate. |
| 10 |
openrouter:stepfun/step-3.7-flash
view ↗ |
8 | 8 | 8 | 8 | 7 | 8 | 7 |
7 |
2.09¢ |
6.1 |
Good SAT-tutor-style questions and analogies, but the rewritten passage is truncated mid-sentence, hurting completeness. Q5 vocab 'surreptitious' not present in the truncated passage. Otherwise solid content. |
| 11 |
openrouter:qwen/qwen3-max-thinking
view ↗ |
8 | 7 | 8 | 8 | 5 | 8 | 8 |
7 |
1.02¢ |
6.7 |
Clean, accurate passage and decent analogies (answer key, counterfeit handbags). Major weakness: only ONE quiz question provided despite the 10-question expectation, sharply reducing quiz value. Glossary and hook strong. |
| 12 |
openrouter:meta-llama/llama-4-maverick
view ↗ |
7 | 6 | 8 | 7 | 7 | 7 | 6 |
7 |
0.26¢ |
7.9 |
Accurate but leans heavily on 'the article says' framing rather than crisp explanation. Analogies serviceable (locked lab, film scouting). Quiz is fair but Q5 'unacceptable' meaning-question is odd. Passage is competent, less engaging. |
| 13 |
openrouter:openai/gpt-5.4-nano
view ↗ |
8 | 6 | 8 | 8 | 8 | 8 | 8 |
7 |
0.63¢ |
7.3 |
Accurate and readable; passage stays close to source with appropriate hedging. Analogies (textbook/locked equipment) are okay but generic. Quiz is fair and answerable. Solid but unremarkable. |
| 14 |
openrouter:amazon/nova-pro-v1
view ↗ |
7 | 5 | 6 | 6 | 6 | 6 | 6 |
6 |
1.34¢ |
6.0 |
Generic chess/doping/homework analogies that don't capture the distillation mechanism precisely. Q1 evidence-pairing references a caption line ('arms race') as though authored. Content is accurate but thin and less engaging; glossary only three terms. |
| 15 |
openrouter:baidu/ernie-4.5-vl-424b-a47b
view ↗ |
7 | 7 | 7 | 7 | 5 | 7 | 7 |
6 |
0.43¢ |
7.2 |
Chef-recipe and counterfeit analogies are fine and grounded. But only 3 quiz questions provided and glossary only 3 terms, reducing depth. Q2 vocab answer 'replicate' over 'condense' is arguable. Passage accurate but compressed. |
Scores 0–10 · 🟩 ≥8 · 🟨 4–7 · 🟥 <4 · sorted by Overall. Hover a column header for its full rubric name.
| # | Model | Quality | Cost | Value |
|---|
| 🥇 |
openrouter:deepseek/deepseek-v4-pro |
9 |
0.62¢ |
8.7 |
| 🥈 |
openrouter:minimax/minimax-m2.7 |
9 |
0.62¢ |
8.7 |
| 🥉 |
openrouter:mistralai/mistral-large-2512 |
8 |
0.85¢ |
8.0 |
| 4 |
openrouter:google/gemini-3.1-flash-lite |
8 |
0.59¢ |
8.0 |
| 5 |
openrouter:meta-llama/llama-4-maverick |
7 |
0.26¢ |
7.9 |
| 6 |
anthropic:claude-haiku-4-5-20251001 |
9 |
2.57¢ |
7.5 |
| 7 |
openrouter:x-ai/grok-4.3 |
8 |
1.64¢ |
7.4 |
| 8 |
openrouter:bytedance-seed/seed-2.0-lite |
8 |
1.73¢ |
7.4 |
| 9 |
openrouter:openai/gpt-5.4-nano |
7 |
0.63¢ |
7.3 |
| 10 |
openrouter:baidu/ernie-4.5-vl-424b-a47b |
6 |
0.43¢ |
7.2 |
| 11 |
openrouter:z-ai/glm-5-turbo |
8 |
3.40¢ |
6.8 |
| 12 |
openrouter:qwen/qwen3-max-thinking |
7 |
1.02¢ |
6.7 |
| 13 |
openrouter:moonshotai/kimi-k2.6 |
9 |
9.62¢ |
6.6 |
| 14 |
openrouter:stepfun/step-3.7-flash |
7 |
2.09¢ |
6.1 |
| 15 |
openrouter:amazon/nova-pro-v1 |
6 |
1.34¢ |
6.0 |
| # | Model | Quality | Cost | Value |
|---|
| 1 |
openrouter:deepseek/deepseek-v4-pro |
9 |
0.62¢ |
8.7 |
| 2 |
openrouter:minimax/minimax-m2.7 |
9 |
0.62¢ |
8.7 |
| 3 |
anthropic:claude-haiku-4-5-20251001 |
9 |
2.57¢ |
7.5 |
| 4 |
openrouter:mistralai/mistral-large-2512 |
8 |
0.85¢ |
8.0 |
| 5 |
openrouter:google/gemini-3.1-flash-lite |
8 |
0.59¢ |
8.0 |
| 6 |
openrouter:x-ai/grok-4.3 |
8 |
1.64¢ |
7.4 |
| 7 |
openrouter:bytedance-seed/seed-2.0-lite |
8 |
1.73¢ |
7.4 |
| 8 |
openrouter:meta-llama/llama-4-maverick |
7 |
0.26¢ |
7.9 |
| 9 |
openrouter:openai/gpt-5.4-nano |
7 |
0.63¢ |
7.3 |
| 10 |
openrouter:qwen/qwen3-max-thinking |
7 |
1.02¢ |
6.7 |
| 11 |
openrouter:stepfun/step-3.7-flash |
7 |
2.09¢ |
6.1 |
| 12 |
openrouter:baidu/ernie-4.5-vl-424b-a47b |
6 |
0.43¢ |
7.2 |
| 13 |
openrouter:amazon/nova-pro-v1 |
6 |
1.34¢ |
6.0 |
Strong quality without paying flagship prices — the cheap-and-good picks.