Science
GPQA Diamond
Graduate-level science questions in biology, physics, and chemistry — hard enough that subject-matter PhDs score around 65%. Higher is better.
Rankings
Higher is better#196%
GPT-6 Astra
294.4%
GPT-5.4-Pro
394.3%
Gemini 3.1 Pro
494.2%
Claude Opus 4.7
494.2%
Kimi K3.1 ProOW
693.6%
GPT-5.5
793.5%
Kimi K3
892.8%
GPT-5.4
992.4%
GPT-5.2
992.4%
Qwen3.7-Max
1191.9%
Gemini 3.0 Pro
1291.7%
Qwen3.8-Flash-NextOW
1391.3%
Claude Opus 4.6
1491.2%
GLM-5.2
1590.9%
DeepSeek-V4.1-FlashOW
1690.5%
Kimi K2.6
1790.4%
Gemini 3.0 Flash
1890.1%
DeepSeek-V4-Pro
1989.9%
Claude Sonnet 4.6
2089.52%
Nemotron 3.5 Super VLOW
2189.5%
Muse Spark
2289.2%
Qwen3.8-27BOW
2388.4%
Qwen3.5OW
2488.1%
GPT-5.1
2488.1%
DeepSeek-V4-Flash
2687.6%
Kimi K2.5
2787%
Claude Opus 4.5
2886.7%
Nemotron 3 UltraOW
2986.43%
Nemotron 3.5 Bolt 74BOW
3086.4%
Gemini 2.5 Pro
3186.2%
GLM-5.1
3286%
GLM-5
3286%
Qwen3.6OW
3485.7%
GLM-4.7
3583.5%
Muse GlimmerOW
3683.4%
Claude Sonnet 4.5
3781%
DeepSeek R1-0528
3880.9%
Claude Opus 4.1
3980.7%
DeepSeek V3.1 Terminus
4080.1%
gpt-oss-120bOW
4179.6%
Claude Opus 4
4279.2%
Nemotron 3 SuperOW
4379.1%
GLM-4.5
4476%
Llama Nemotron Ultra 253BOW
4575.7%
o1
4675.4%
Claude Sonnet 4
4675.4%
Nemotron 3.5 LightningOW
4875.1%
Kimi K2
4875.1%
Kimi K2 (0905)
5073.3%
o1-preview
5173%
Claude Haiku 4.5
5173%
Nemotron 3 NanoOW
5371.5%
gpt-oss-20bOW
5468%
Claude 3.7 Sonnet
5564%
Nemotron Nano 2OW
5660%
o1-mini
5759.4%
Claude 3.5 Sonnet
5859.1%
DeepSeek V3 Chat
5950.4%
Claude 3 Opus
6049.9%
GPT-4o
6145.96%
Mistral Small 3.1
6242.5%
GPT-4 Turbo
6342.4%
Gemma 3OW
6441.6%
Claude 3.5 Haiku
6540.4%
Claude 3 Sonnet
6640.2%
GPT-4o mini
6733.3%
Claude 3 Haiku
GPQA Diamond — frequently asked questions
Graduate-level science questions in biology, physics, and chemistry — hard enough that subject-matter PhDs score around 65%. Higher is better.