Science Reasoning on GPQA Diamond (Accuracy (%))
75.7AccuracyDeepSeek-R1
Evaluation Results
| Method | Links | |
|---|---|---|
| DeepSeek-R12025.08 | 75.7 | |
| DS-32B-MathDomain=Math2025.08 | 68.69 | |
| NTele-R1-32B2025.08 | 67.68 | |
| DS-32B-CodeDomain=Code2025.08 | 65.15 | |
| QwQ-32B2025.08 | 63.6 | |
| DS-32B2025.08 | 62.1 |