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 better194.4%
GPT-5.4-Pro
OpenAI · Mar 5 2026
294.3%
Gemini 3.1 Pro
Google · Feb 19 2026
394.2%
Claude Opus 4.7
Anthropic · Apr 16 2026
493.6%
GPT-5.5
OpenAI · Apr 23 2026
593.5%
Kimi K3
Moonshot AI · Jul 16 2026
692.8%
GPT-5.4
OpenAI · Mar 5 2026
792.4%
GPT-5.2
OpenAI · Dec 11 2025
891.9%
Gemini 3.0 Pro
Google · Nov 18 2025
991.3%
Claude Opus 4.6
Anthropic · Feb 5 2026
1091.2%
GLM-5.2
Z.ai · Jun 16 2026
1190.5%
Kimi K2.6
Moonshot AI · Apr 21 2026
1290.4%
Gemini 3.0 Flash
Google · Dec 17 2025
1390.1%
DeepSeek-V4-Pro
DeepSeek · Apr 24 2026
1489.9%
Claude Sonnet 4.6
Anthropic · Feb 17 2026
1589.5%
Muse Spark
Meta · Apr 8 2026
1688.1%
GPT-5.1
OpenAI · Nov 12 2025
1688.1%
DeepSeek-V4-Flash
DeepSeek · Apr 24 2026
1887.6%
Kimi K2.5
Moonshot AI · Jan 27 2026
1987%
Claude Opus 4.5
Anthropic · Nov 24 2025
2086.4%
Gemini 2.5 Pro
Google · Mar 25 2025
2186.2%
GLM-5.1
Z.ai · Apr 7 2026
2286%
GLM-5
Z.ai · Feb 12 2026
2385.7%
GLM-4.7
Z.ai · Dec 22 2025
2483.4%
Claude Sonnet 4.5
Anthropic · Sep 29 2025
2580.9%
Claude Opus 4.1
Anthropic · Aug 5 2025
2679.6%
Claude Opus 4
Anthropic · May 22 2025
2779.1%
GLM-4.5
Z.ai · Jul 28 2025
2875.4%
Claude Sonnet 4
Anthropic · May 22 2025
2975.1%
Kimi K2
Moonshot AI · Jul 11 2025
2975.1%
Kimi K2 (0905)
Moonshot AI · Sep 5 2025
3173%
Claude Haiku 4.5
Anthropic · Oct 15 2025
3268%
Claude 3.7 Sonnet
Anthropic · Feb 24 2025
3359.4%
Claude 3.5 Sonnet
Anthropic · Jun 20 2024
3450.4%
Claude 3 Opus
Anthropic · Mar 4 2024
3541.6%
Claude 3.5 Haiku
Anthropic · Oct 22 2024
3640.4%
Claude 3 Sonnet
Anthropic · Mar 4 2024
3733.3%
Claude 3 Haiku
Anthropic · Mar 4 2024