Best LLMs for Science — Qwen
The best Qwen model for science is Qwen3.7 Max, scoring 92.3% on GPQA Diamond — graduate-level physics, chemistry and biology. Qwen3.7 Plus (90.0%) and Qwen3.5 397B A17B (89.3%) round out the top three.
- 1
qwen3.7-maxReasoningToolsJSON46.0 intel · $1.48/M · 44 t/s92.3%GPQA - 2qwen3.7-plusReasoningToolsJSON+139.0 intel · $0.320/M · 12 t/s90.0%GPQA
- 3qwen3.5-397b-a17bReasoningToolsJSON+133.7 intel · $0.390/M · 69 t/s89.3%GPQA
- 4qwen3.6-max-previewReasoningToolsJSON40.0 intel · $1.03/M · 53 t/s88.8%GPQA
- 5qwen3.6-plusReasoningToolsJSON+139.6 intel · $0.325/M · 37 t/s88.2%GPQA
- 6qwen3-max-thinkingReasoningToolsJSON31.7 intel · $0.780/M · 262K ctx86.1%GPQA
- 7qwen3.5-27bReasoningToolsJSON+133.8 intel · $0.195/M · 38 t/s85.8%GPQA
- 8qwen3.5-122b-a10bReasoningToolsJSON+132.3 intel · $0.260/M · 130 t/s85.7%GPQA
- 9qwen3.5-35b-a3bReasoningToolsJSON+129.3 intel · $0.140/M · 217 t/s84.5%GPQA
- 10qwen3.6-27bReasoningToolsJSON+137.1 intel · $0.289/M · 50 t/s84.2%GPQA
- 11qwen3.6-35b-a3bReasoningToolsJSON+131.6 intel · $0.140/M · 154 t/s84.1%GPQA
- 12qwen3.5-9bReasoningToolsJSON+121.4 intel · $0.100/M · 70 t/s80.6%GPQA
- 13qwen3-maxToolsJSON24.0 intel · $0.780/M · 262K ctx76.4%GPQA
- 14qwen3-next-80b-a3b-instructToolsJSON13.7 intel · $0.090/M · 205 t/s73.8%GPQA
- 15qwen3-coder-nextToolsJSON21.1 intel · $0.120/M · 133 t/s73.7%GPQA
- 16qwen3-vl-235b-a22b-instructToolsJSONVision14.3 intel · $0.210/M · 262K ctx71.2%GPQA
- 17qwen3-vl-30b-a3b-instructToolsJSONVision10.0 intel · $0.150/M · 262K ctx69.5%GPQA
- 18qwen3-vl-32b-instructToolsJSONVision11.1 intel · $0.104/M · 131K ctx67.1%GPQA
- 19qwen3-coder-30b-a3b-instructToolsJSON13.6 intel · $0.070/M · 81 t/s51.6%GPQA
- 20qwen-2.5-72b-instructToolsJSON9.6 intel · $0.360/M · 24 t/s49.1%GPQA
- 21qwen3-vl-8b-instructToolsJSONVision8.4 intel · $0.117/M · 262K ctx42.7%GPQA
- 22qwen-2.5-coder-32b-instruct7.1 intel · $0.660/M · 16 t/s41.7%GPQA
How this is ranked
AI models ranked by GPQA Diamond — graduate-level physics, chemistry and biology questions written to be un-Googleable, so the score reflects reasoning from knowledge rather than retrieval. The best large language models for scientific research and hard-science work.
Frequently asked
What is the best Qwen model for science?
The best Qwen model for science is Qwen3.7 Max, scoring 92.3% on GPQA Diamond — graduate-level physics, chemistry and biology. Qwen3.7 Plus (90.0%) and Qwen3.5 397B A17B (89.3%) round out the top three.
What's a good alternative to Qwen3.7 Max?
Qwen3.7 Plus (90.0%) is the closest alternative on this metric, followed by Qwen3.5 397B A17B (89.3%). See the full ranking above for the tradeoffs.
How many Qwen models are there?
modelgrep tracks 49 Qwen models with live benchmarks, speed, latency and per-provider pricing, led on intelligence by Qwen3.7 Max. 22 of them qualify for this ranking.