public · science
GPQA Diamond
Graduate-level science questions designed to be Google-proof.
Items
198
Revision
2024.11
Contamination
watched
Scorer
scorer.choice.v3
- Dataset revision
- [email protected]
- Scorer
- exact-choice · scorer.choice.v3
- License
- CC BY 4.0
- Kind
- Public reproduction suite
| Model | Score | $/task | Median s | Tools | Evidence |
|---|---|---|---|---|---|
| o4-pro | 90.4 | $0.638 | 38.2 | 0.2 | LIVE_MEASURED |
| Grok 4 Heavy | 88.1 | $0.132 | 22.6 | 0.2 | LIVE_MEASURED |
| Grok 4 | 84.8 | $0.038 | 10.4 | 0.2 | LIVE_MEASURED |
| Gemini 2.5 Pro | 83.4 | $0.025 | 8.7 | 0.2 | LIVE_MEASURED |
| GPT-5 | 82.6 | $0.063 | 12.9 | 0.2 | LIVE_MEASURED |
| Claude 4.1 Opus | 81.2 | $0.267 | 16.8 | 0.2 | LIVE_MEASURED |
| DeepSeek V3.1 | 79.6 | $0.0042 | 8.2 | 0.2 | REPRODUCED |
| Qwen3-Max | 78.2 | $0.0056 | 9.7 | 0.2 | LIVE_MEASURED |
| Claude 4.1 Sonnet | 76.8 | $0.047 | 11.4 | 0.2 | LIVE_MEASURED |
| Mistral Large 3 | 74.6 | $0.015 | 9.9 | 0.2 | LIVE_MEASURED |
| Gemini 2.5 Flash | 72.1 | $0.0011 | 3.3 | 0.2 | LIVE_MEASURED |
| GPT-5 Mini | 71.4 | $0.0031 | 5.1 | 0.2 | LIVE_MEASURED |
| Llama 4 405B | 70.8 | $0.0043 | 9.1 | 0.2 | REPRODUCED |
| Grok 3 Mini | 68.2 | $0.0016 | 4.0 | 0.2 | LIVE_MEASURED |
| Llama 4 70B | 61.4 | $0.00094 | 1.9 | 0.2 | REPRODUCED |