← Artificial Analysis leaderboards
GPQA Diamond
Artificial Analysis · 50 models
| Rank | Model | Score |
|---|---|---|
| 1 | GPT-6 Astra (Xhigh) | 0.963 |
| 2 | GPT-6 Astra (Max) | 0.961 |
| 3 | Gemini 3.8 Flash (High) | 0.953 |
| 4 | GPT-6 Astra (High) | 0.949 |
| 4 | Grok 4.6 (High) | 0.949 |
| 6 | Gemini 3.7 Flash (High) | 0.945 |
| 7 | GPT-5.6 Sol (Max) | 0.941 |
| 7 | Gemini 3.1 Pro Preview | 0.941 |
| 7 | Muse Spark 1.3 (Xhigh) | 0.941 |
| 10 | GPT-6 Astra (Medium) | 0.939 |
| 11 | Claude Fable 5.1 (Max, Default Fallback) | 0.937 |
| 11 | Claude Opus 5 (High) | 0.937 |
| 11 | Claude Opus 5 (Xhigh) | 0.937 |
| 14 | GPT-5.5 (Xhigh) | 0.935 |
| 14 | Gemini 3.8 Flash (Medium) | 0.935 |
| 14 | Grok 4.6 (Medium) | 0.935 |
| 14 | Grok 4.6 (Xhigh) | 0.935 |
| 14 | Kimi K3 (Max) | 0.935 |
| 14 | Muse Spark 1.3 (Max) | 0.935 |
| 14 | Qwen3.8 2.4T A95B | 0.935 |
| 21 | Claude Fable 5.1 (Xhigh, Default Fallback) | 0.934 |
| 22 | Claude Opus 5 (Max) | 0.932 |
| 22 | GPT-5.5 (High) | 0.932 |
| 24 | GPT-5.6 Sol (Xhigh) | 0.931 |
| 24 | GPT-6 Astra (Low) | 0.931 |
| 24 | Grok 4.5 (High) | 0.931 |
| 27 | MiniMax-M3 | 0.929 |
| 28 | DeepSeek V4 Pro 0813 (Max) | 0.928 |
| 28 | GPT-5.6 Sol (High) | 0.928 |
| 28 | Gemini 3.6 Flash (High) | 0.928 |
| 28 | Qwen3.8 Max (0902) | 0.928 |
| 32 | Qwen3.8 Max | 0.927 |
| 33 | Claude Fable 5 (Max, Opus 4.8 Fallback) | 0.926 |
| 33 | GPT-5.5 (Medium) | 0.926 |
| 33 | GPT-5.6 Sol (Medium) | 0.926 |
| 36 | GPT-5.6 Terra (Max) | 0.925 |
| 37 | Qwen3.7 Max | 0.923 |
| 37 | Qwen3.8-Flash-Next | 0.923 |
| 39 | Gemini 3.5 Flash (High) | 0.922 |
| 40 | Gemini 3.5 Flash (Medium) | 0.921 |
| 40 | Gemini 3.7 Flash (Medium) | 0.921 |
| 42 | Claude Opus 4.8 (Max) | 0.92 |
| 42 | GPT-5.4 (Xhigh) | 0.92 |
| 42 | Gemini 3.8 Flash (Low) | 0.92 |
| 45 | Claude Opus 5 (Medium) | 0.919 |
| 46 | GLM-5.3 (Max) | 0.917 |
| 47 | GPT-5.3 Codex (Xhigh) | 0.915 |
| 48 | Claude Opus 4.7 (Max) | 0.914 |
| 49 | DeepSeek V4 Flash Vision (Max) | 0.913 |
| 50 | GLM-5.3-Flash | 0.912 |