GPQA Diamond
GPQA Diamond - Google-Proof Q&A benchmark for graduate-level scientific reasoning.
Category
STEM & Reasoning
Max Score
100.0
Score Type
percent
Active
Yes
Model Scores
Scores for GPQA Diamond
| Model | Score | Date | Verified |
|---|---|---|---|
| Qwen3.6-35B-A3B |
84.31%
|
16.04.2026 | Verified |
| Gemma4-31B-it |
81.09%
|
16.04.2026 | Unverified |
| Qwen3.5-35B-A3B |
80.90%
|
16.04.2026 | Unverified |
| Gemma4-26B-A4B |
77.31%
|
16.04.2026 | Unverified |
| Qwen3.6-27B |
87.71%
|
22.04.2026 | Verified |
| Qwen3.5-397B-A17B |
88.84%
|
22.04.2026 | Unverified |
| Claude 4.5 Opus |
86.20%
|
22.04.2026 | Unverified |
| Qwen3.5-27B |
83.36%
|
16.04.2026 | Unverified |
| NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16 |
86.20%
|
04.06.2026 | Verified |
| MiniMax-M2.7 |
85.44%
|
04.06.2026 | Unverified |
| GLM-5.1 |
84.50%
|
04.06.2026 | Unverified |
| Kimi K2.6 |
93.76%
|
04.06.2026 | Unverified |
| DeepSeek-V4-Pro |
87.71%
|
04.06.2026 | Unverified |
| Qwen3.8-27B |
90.36%
|
18.08.2026 | Verified |
| Qwen3.7-Plus |
92.44%
|
18.08.2026 | Unverified |
| Muse Glimmer-30B |
79.58%
|
18.08.2026 | Unverified |
| Opus4.6 Max |
94.33%
|
18.08.2026 | Unverified |
| Fable-5 |
96.79%
|
18.08.2026 | Unverified |
| GPT-5.6 Sol |
99.62%
|
18.08.2026 | Unverified |
| Claude Opus 4.8 |
95.65%
|
18.08.2026 | Unverified |
| Qwen3.8-2.4T-A95B |
96.79%
|
18.08.2026 | Verified |
| Qwen3.7-Max |
91.87%
|
17.06.2026 | Unverified |
| Qwen3.5-4B |
65.78%
|
01.02.2026 | Verified |
| Muse-Glimmer-30B |
79.58%
|
01.08.2026 | Verified |
| Kimi K3 |
98.49%
|
18.08.2026 | Verified |
| GLM-5.2 |
94.14%
|
17.06.2026 | Verified |
| Gemini 3.1 Pro |
100.00%
|
17.06.2026 | Unverified |
| GPT-5.5 |
98.68%
|
17.06.2026 | Unverified |
| MiniMax-M3 |
97.54%
|
17.06.2026 | Unverified |
| Qwen3.5-122B-A10B |
85.44%
|
01.02.2026 | Verified |
| Qwen3-235B-A22B |
75.04%
|
01.02.2026 | Unverified |
| GPT-5-mini 2025-08-07 |
78.26%
|
01.02.2026 | Unverified |
| Gemma4-12B |
70.69%
|
02.07.2026 | Verified |
| Gemma4-E4B |
32.50%
|
02.07.2026 | Verified |
| Gemma4-E2B |
3.76%
|
02.07.2026 | Verified |
| Gemma3-27B |
1.87%
|
02.07.2026 | Unverified |
| Ornith-1.5-35B-A3B |
90.36%
|
19.08.2026 | Verified |
| Ornith-1.0-35B-A3B |
84.69%
|
19.08.2026 | Unverified |
| Gemma-4-31B |
81.09%
|
19.08.2026 | Unverified |
| GPT-5.4 |
97.16%
|
23.08.2026 | Verified |
| Claude Opus 4.6 |
94.33%
|
23.08.2026 | Verified |
| Kimi K2.5 |
87.33%
|
23.08.2026 | Verified |
| GPT-OSS-120B |
73.15%
|
08.08.2025 | Verified |
| GPT-OSS-20B |
56.89%
|
05.08.2025 | Verified |
| Gemini-2.5-Flash-Thinking |
78.26%
|
31.07.2025 | Verified |
| Qwen3-30B-A3B |
46.11%
|
29.04.2025 | Verified |
| Qwen3-30B-A3B-Thinking-2507 |
60.48%
|
31.07.2025 | Verified |
| DeepSeek-V4-Flash |
88.28%
|
26.06.2026 | Verified |
| Opus-4.6 |
94.33%
|
26.06.2026 | Verified |
| DeepSeek-V4-Pro |
92.06%
|
26.06.2026 | Verified |
| Gemma 3 27B IT |
41.41
|
27.02.2025 | Verified |
| Qwen3.8-Flash-Next |
95.08%
|
26.08.2026 | Verified |
| DeepSeek-V4-Flash-0731 |
93.38%
|
26.08.2026 | Unverified |
| Claude-Opus-4.6 (Max) |
94.33%
|
26.08.2026 | Unverified |
| Hy3 |
93.57%
|
28.08.2026 | Verified |
| DeepSeek V4 Pro |
97.16%
|
28.08.2026 | Verified |
| GLM-5.3 |
95.08%
|
28.08.2026 | Verified |
| Claude Opus 5 |
98.87%
|
28.08.2026 | Verified |
| MiniCPM5-2B |
54.43%
|
— | Verified |
| LFM2.5-2.6B |
27.21%
|
— | Verified |
| Qwen3.5-2B |
7.92%
|
— | Verified |
| granite-4.2-3B |
27.40%
|
— | Verified |
| Nemotron-3-Nano-4B |
18.70%
|
— | Verified |
| LFM2.5-8B-A1B |
18.70%
|
— | Verified |
| DeepSeek-V4.1-Flash |
93.57%
|
10.09.2026 | Verified |
| Hy4 preview |
96.22%
|
28.08.2026 | Verified |