Scientific Reasoning on GPQA Diamond (Domain Metrics)
87.37Average ScoreInternAgent-1.5
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| InternAgent-1.5Model Category=Our Method2026.02 | 87.37 | 84.21 | 79.57 | 96.51 | |
| GPT-5Model Category=Base Models2026.02 | 85.35 | 84.21 | 76.34 | 95.35 | |
| MiroThinkerModel Category=React Model with Tools2026.02 | 84.85 | 84.21 | 75.27 | 95.35 | |
| Deepseek-R1Model Category=Base Models2026.02 | 82.32 | 63.16 | 76.34 | 91.86 | |
| Tongyi DRModel Category=React Model with Tools2026.02 | 80.3 | 78.95 | 67.74 | 95.35 | |
| o4-miniModel Category=Base Models2026.02 | 78.28 | 78.95 | 63.44 | 94.19 | |
| Intern-S1Model Category=Base Models2026.02 | 78.26 | 89.47 | 59.49 | 93.02 | |
| WebShaperModel Category=React Model with Tools2026.02 | 64.65 | 47.37 | 52.69 | 81.4 | |
| Qwen3-32BModel Category=Base Models2026.02 | 49.49 | — | — | — | |
| Qwen3-235BModel Category=Base Models2026.02 | 47.47 | — | — | — | |
| Qwen-3-8BModel Category=Base Models2026.02 | 44.44 | — | — | — |