GPQA Diamond (Epoch)
Epoch AI · source page ↗ · last checked Aug 24, 2026, 12:02 PM
Current leaderboard
captured Aug 24, 2026, 12:02 PM| Model | GPQA Diamond accuracy (0–1) |
|---|---|
| gemini-3.7-flash_high | 0.948 |
| gpt-5.4-pro-2026-03-05_xhigh | 0.946 |
| gemini-3.1-pro-preview_high | 0.944 |
| gemini-3.6-flash_high | 0.941 |
| gemini-3.1-pro-preview | 0.941 |
| grok-4.6_high | 0.94 |
| gpt-5.5-pre-release_xhigh | 0.94 |
| claude-opus-5_max | 0.939 |
| gpt-5.5-pro-pre-release_xhigh | 0.939 |
| gpt-5.6-sol_max | 0.935 |
| grok-4.5_high | 0.934 |
| gpt-5.6-terra_max | 0.933 |
| gpt-5.4-2026-03-05_xhigh | 0.933 |
| grok-4.6_xhigh | 0.932 |
| kimi-k3_max | 0.931 |
| claude-opus-5 | 0.929 |
| gemini-3.5-flash_high | 0.928 |
| qwen3.8-max_xhigh | 0.927 |
| gemini-3-pro-preview | 0.926 |
| kimi-k3_high | 0.919 |
| glm-5.2_max | 0.919 |
| deepseek-v4-pro-0813_max | 0.917 |
| gpt-5.6-luna_max | 0.916 |
| gpt-5.2-2025-12-11_xhigh | 0.914 |
| deepseek-v4-flash-0731_max | 0.91 |
Change history
- Benchmark Aug 20, 2026, 12:02 AM
GPQA Diamond (Epoch) scores changed
GPQA Diamond accuracy (0–1) updated.
View raw diff +1 −0
+ deepseek-v4-pro-0813_max 0.917 - Benchmark Aug 16, 2026, 12:02 AM
GPQA Diamond (Epoch) scores changed
GPQA Diamond accuracy (0–1) updated.
View raw diff +3 −0
+ gemini-3.7-flash_high 0.948 + grok-4.6_xhigh 0.932 + inkling-small_xhigh 0.885
- Benchmark Aug 14, 2026, 12:02 AM
GPQA Diamond (Epoch) scores changed
GPQA Diamond accuracy (0–1) updated.
View raw diff +3 −0
+ gpt-5.5-instant 0.825 + grok-4.6_high 0.94 + Inkling_xhigh 0.883
- Silent update Aug 11, 2026, 12:02 AM
glm-5.1: GPQA Diamond accuracy (0–1) ↑ 0.855 → 0.899
Same model id, score moved on GPQA Diamond (Epoch) — a silent re-evaluation or model swap.
- Benchmark Aug 11, 2026, 12:02 AM
GPQA Diamond (Epoch) scores changed
GPQA Diamond accuracy (0–1) updated.
View raw diff +3 −1
- glm-5.1 0.855 + glm-5.1 0.899 + glm-5.2_none 0.712 + MiniMax-M3 0.909
- Silent update Aug 7, 2026, 6:02 PM
gemma-3-27b-it: GPQA Diamond accuracy (0–1) ↓ 0.489 → 0.439
Same model id, score moved on GPQA Diamond (Epoch) — a silent re-evaluation or model swap.
- Silent update Aug 7, 2026, 6:02 PM
magistral-small-2506: GPQA Diamond accuracy (0–1) ↑ 0.484 → 0.561
Same model id, score moved on GPQA Diamond (Epoch) — a silent re-evaluation or model swap.
- Benchmark Aug 7, 2026, 6:02 PM
GPQA Diamond (Epoch) scores changed
GPQA Diamond accuracy (0–1) updated.
View raw diff +11 −2
- gemma-3-27b-it 0.489 - magistral-small-2506 0.484 + gemma-3-27b-it 0.439 + gpt-oss-20b_high 0.46 + magistral-small-2506 0.561 + qwen3.5-397b-a17b 0.859 + qwen3.5-397b-a17b_none 0.864 + qwen3.6-27b 0.859 + qwen3.6-27b_none 0.848 + qwen3.6-35b-a3b_none 0.848 + qwen3.7-flash 0.823 + qwen3.7-flash_none 0.808 + qwen3.7-plus_none 0.818
- Silent update Aug 7, 2026, 12:02 PM
qwen3.7-max: GPQA Diamond accuracy (0–1) ↓ 0.916 → 0.909
Same model id, score moved on GPQA Diamond (Epoch) — a silent re-evaluation or model swap.
- Benchmark Aug 7, 2026, 12:02 PM
GPQA Diamond (Epoch) scores changed
GPQA Diamond accuracy (0–1) updated.
View raw diff +3 −1
- qwen3.7-max 0.916 + deepseek-v4-pro_high 0.909 + qwen3.7-max 0.909 + qwen3.7-plus 0.879
- Silent update Aug 7, 2026, 6:02 AM
qwen3.5-flash: GPQA Diamond accuracy (0–1) ↓ 0.838 → 0.823
Same model id, score moved on GPQA Diamond (Epoch) — a silent re-evaluation or model swap.
- Silent update Aug 7, 2026, 6:02 AM
qwen3.6-flash: GPQA Diamond accuracy (0–1) ↓ 0.844 → 0.833
Same model id, score moved on GPQA Diamond (Epoch) — a silent re-evaluation or model swap.
- Silent update Aug 7, 2026, 6:02 AM
qwen3.5-plus: GPQA Diamond accuracy (0–1) ↑ 0.842 → 0.848
Same model id, score moved on GPQA Diamond (Epoch) — a silent re-evaluation or model swap.
- Silent update Aug 7, 2026, 6:02 AM
qwen3.6-max-preview: GPQA Diamond accuracy (0–1) ↓ 0.891 → 0.874
Same model id, score moved on GPQA Diamond (Epoch) — a silent re-evaluation or model swap.
- Silent update Aug 7, 2026, 6:02 AM
kimi-k2.7-code: GPQA Diamond accuracy (0–1) ↓ 0.895 → 0.879
Same model id, score moved on GPQA Diamond (Epoch) — a silent re-evaluation or model swap.
- Silent update Aug 7, 2026, 6:02 AM
qwen3.6-plus: GPQA Diamond accuracy (0–1) ↑ 0.874 → 0.884
Same model id, score moved on GPQA Diamond (Epoch) — a silent re-evaluation or model swap.
- Benchmark Aug 7, 2026, 6:02 AM
GPQA Diamond (Epoch) scores changed
GPQA Diamond accuracy (0–1) updated.
View raw diff +45 −6
- kimi-k2.7-code 0.895 - qwen3.5-flash 0.838 - qwen3.5-plus 0.842 - qwen3.6-flash 0.844 - qwen3.6-max-preview 0.891 - qwen3.6-plus 0.874 + claude-fable-5_high 0.833 + claude-fable-5_max 0.859 + claude-opus-4-6_max 0.884 + claude-opus-4-7_max 0.864 + claude-opus-4-8_low 0.884 + claude-opus-5 0.929 + claude-sonnet-4-6_max 0.788 + claude-sonnet-5_max 0.803 + deepseek-v4-pro_none 0.732 + gemini-3-flash-preview_high 0.894 + gemini-3.1-flash-lite_high 0.818 + gemini-3.1-flash-lite_low 0.742 + gemini-3.1-flash-lite_minimal 0.737 + gemini-3.1-pro-preview_high 0.944 + gemini-3.5-flash_low 0.889 + gemini-3.5-flash-lite_high 0.833 + gemini-3.5-flash-lite_low 0.758 + gemini-3.5-flash-lite_minimal 0.742 + gemini-3.6-flash_low 0.864 + gemini-3.6-flash_minimal 0.859 + gemma-4-31b-it_minimal 0.758 + gpt-5-mini-2025-08-07_minimal 0.717 + gpt-5.1-2025-11-13_none 0.667 + gpt-5.4-mini-2026-03-17_none 0.641 + gpt-5.4-mini-2026-03-17_xhigh 0.869 + gpt-5.4-nano-2026-03-17_low 0.722 + gpt-5.4-nano-2026-03-17_none 0.556 + gpt-5.5_none 0.773 + gpt-5.6-luna_low 0.823 + gpt-5.6-luna_none 0.636 + gpt-5.6-sol_low 0.899 + gpt-5.6-sol_none 0.828 + gpt-5.6-terra_low 0.874 + gpt-5.6-terra_none 0.773 + kimi-k2.7-code 0.879 + kimi-k3_high 0.919 + kimi-k3_low 0.848 + o3-2025-04-16_medium 0.808 + o4-mini-2025-04-16_medium 0.778 + qwen3.5-flash 0.823 + qwen3.5-plus 0.848 + qwen3.6-35b-a3b 0.838 + qwen3.6-flash 0.833 + qwen3.6-max-preview 0.874 + qwen3.6-plus 0.884
- Benchmark Aug 7, 2026, 12:02 AM
GPQA Diamond (Epoch) scores changed
GPQA Diamond accuracy (0–1) updated.
View raw diff +3 −0
+ claude-fable-5_low 0.788 + claude-opus-4-8_none 0.854 + claude-opus-5_low 0.879
- Benchmark Aug 5, 2026, 6:02 PM
GPQA Diamond (Epoch) scores changed
GPQA Diamond accuracy (0–1) updated.
View raw diff +1 −0
+ qwen3.8-max_xhigh 0.927 - Benchmark Aug 3, 2026, 12:02 AM
GPQA Diamond (Epoch) scores changed
GPQA Diamond accuracy (0–1) updated.
View raw diff +1 −0
+ deepseek-v4-flash-0731_max 0.91 - Benchmark Aug 2, 2026, 6:01 AM
GPQA Diamond (Epoch) scores changed
GPQA Diamond accuracy (0–1) updated.
View raw diff +1 −0
+ gemini-3.6-flash_high 0.941 - Benchmark Jul 29, 2026, 12:02 AM
GPQA Diamond (Epoch) scores changed
GPQA Diamond accuracy (0–1) updated.
View raw diff +16 −0
+ claude-sonnet-4-6_high 0.833 + claude-sonnet-4-6_medium 0.833 + deepseek-chat 0.712 + gemini-3.5-flash_minimal 0.864 + gpt-5-2025-08-07_minimal 0.717 + gpt-5-nano-2025-08-07_low 0.576 + gpt-5-nano-2025-08-07_minimal 0.485 + gpt-5.2-2025-12-11_none 0.732 + gpt-5.4-2026-03-05_high 0.899 + gpt-5.4-2026-03-05_low 0.848 + gpt-5.4-2026-03-05_medium 0.889 + gpt-5.4-2026-03-05_none 0.747 + o1-2024-12-17_low 0.742 + o3-2025-04-16_low 0.798 + o3-mini-2025-01-31_low 0.682 + o4-mini-2025-04-16_low 0.753
- Benchmark Jul 25, 2026, 12:02 AM
GPQA Diamond (Epoch) scores changed
GPQA Diamond accuracy (0–1) updated.
View raw diff +1 −0
+ claude-opus-5_max 0.939 - Benchmark Jul 18, 2026, 12:02 AM
GPQA Diamond (Epoch) scores changed
GPQA Diamond accuracy (0–1) updated.
View raw diff +1 −0
+ kimi-k3_max 0.931 - Benchmark Jul 13, 2026, 6:02 PM
GPQA Diamond (Epoch) scores changed
GPQA Diamond accuracy (0–1) updated.
View raw diff +1 −0
+ grok-4.20-0309-reasoning 0.893 - Benchmark Jul 10, 2026, 12:02 AM
GPQA Diamond (Epoch) scores changed
GPQA Diamond accuracy (0–1) updated.
View raw diff +5 −0
+ claude-sonnet-5_xhigh 0.905 + gpt-5.6-luna_max 0.916 + gpt-5.6-sol_max 0.935 + gpt-5.6-terra_max 0.933 + grok-4.5_high 0.934
- Benchmark Jun 25, 2026, 4:21 PM
GPQA Diamond (Epoch) scores changed
GPQA Diamond accuracy (0–1) updated.
View raw diff +1 −0
+ glm-5.2_max 0.919