Best LLMs for Science — OpenAI
The best OpenAI model for science is GPT-5.6 Sol, scoring 94.1% on GPQA Diamond — graduate-level physics, chemistry and biology. GPT-5.5 (93.5%) and GPT-5.6 Terra (92.5%) round out the top three.
- 1
gpt-5.6-solReasoningToolsJSON+158.9 intel · $5.00/M · 51 t/s94.1%GPQA - 2gpt-5.5ReasoningToolsJSON+154.8 intel · $5.00/M · 1.1M ctx93.5%GPQA
- 3gpt-5.6-terraReasoningToolsJSON+155.0 intel · $1.00/M · 83 t/s92.5%GPQA
- 4gpt-5.4ReasoningToolsJSON+151.4 intel · $2.50/M · 1.1M ctx92.0%GPQA
- 5gpt-5.3-codexReasoningToolsJSON+144.3 intel · $1.75/M · 122 t/s91.5%GPQA
- 6gpt-5.6-lunaReasoningToolsJSON+151.2 intel · $0.100/M · 111 t/s91.1%GPQA
- 7gpt-5.2ReasoningToolsJSON+142.2 intel · $1.75/M · 81 t/s90.3%GPQA
- 8gpt-5.2-codexReasoningToolsJSON+140.1 intel · $1.75/M · 400K ctx89.9%GPQA
- 9gpt-5.4-miniReasoningToolsJSON+140.0 intel · $0.750/M · 400K ctx87.5%GPQA
- 10gpt-5.1ReasoningToolsJSON+136.9 intel · $1.25/M · 59 t/s87.3%GPQA
- 11gpt-5.1-codexReasoningToolsJSON+134.7 intel · $1.25/M · 65 t/s86.0%GPQA
- 12gpt-5ReasoningToolsJSON+134.7 intel · $1.25/M · 65 t/s85.4%GPQA
- 13o3-proReasoningToolsJSON+132.5 intel · $20.00/M · 3 t/s84.5%GPQA
- 14gpt-5-miniReasoningToolsJSON+125.3 intel · $0.250/M · 93 t/s82.8%GPQA
- 15o3ReasoningToolsJSON+130.4 intel · $2.00/M · 54 t/s82.7%GPQA
- 16gpt-5.4-nanoReasoningToolsJSON+138.2 intel · $0.200/M · 400K ctx81.7%GPQA
- 17gpt-5.1-codex-miniReasoningToolsJSON+130.6 intel · $0.250/M · 61 t/s81.3%GPQA
- 18o4-mini-highReasoningToolsJSON+125.6 intel · $1.10/M · 84 t/s78.4%GPQA
- 19o4-miniReasoningToolsJSON+125.6 intel · $1.10/M · 59 t/s78.4%GPQA
- 20gpt-oss-120bReasoningToolsJSON23.8 intel · $0.037/M · 624 t/s78.2%GPQA
- 21o3-mini-highReasoningToolsJSON15.6 intel · $1.10/M · 200K ctx77.3%GPQA
- 22o3-miniReasoningToolsJSON19.0 intel · $1.10/M · 200K ctx74.8%GPQA
- 23o1ReasoningToolsJSON+123.4 intel · $15.00/M · 200K ctx74.7%GPQA
- 24gpt-oss-20bReasoningToolsJSON14.9 intel · $0.030/M · 236 t/s68.8%GPQA
- 25gpt-oss-20b:freeReasoningToolsJSON14.9 intel · Free/M · 17 t/s68.8%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 OpenAI model for science?
The best OpenAI model for science is GPT-5.6 Sol, scoring 94.1% on GPQA Diamond — graduate-level physics, chemistry and biology. GPT-5.5 (93.5%) and GPT-5.6 Terra (92.5%) round out the top three.
What's a good alternative to GPT-5.6 Sol?
GPT-5.5 (93.5%) is the closest alternative on this metric, followed by GPT-5.6 Terra (92.5%). See the full ranking above for the tradeoffs.
How many OpenAI models are there?
modelgrep tracks 60 OpenAI models with live benchmarks, speed, latency and per-provider pricing, led on intelligence by GPT-5.6 Sol. 25 of them qualify for this ranking.