Mathematical and Scientific Reasoning on Combined Suite (AIME, HMMT, GPQA, MMLU-Pro, MMLU-Redux 2.0)
89.5Pass@1GPT-OSS-120B
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| GPT-OSS-120B2026.04 | 89.5 | 3,661.45 | 24.44 | |
| Qwen3-30B-A3BMode=Thinking2026.04 | 89.42 | 4,299.45 | 20.8 | |
| Gemini 2.5 Flash2026.04 | 89.13 | 4,137.27 | 21.54 | |
| GPT-5 Mini2026.04 | 88.85 | 3,616.77 | 24.57 | |
| Aryabhata 22026.04 | 87.64 | 2,214.35 | 39.58 | |
| GPT-OSS-20B2026.04 | 84.95 | 4,859.79 | 17.48 | |
| Nemotron 3 Nano 30B A3B2026.04 | 83.48 | 6,132.9 | 13.61 | |
| GPT-5 Nano2026.04 | 79.51 | 5,222.71 | 15.22 |