Mathematical Reasoning on AIME 25 (pass@4, pass@8, pass@16)
32.64Pass@4Qwen2.5-7B EP-GRPO
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Qwen2.5-7B EP-GRPOModel=Qwen2.5-7B, Method=EP-GRPO2026.05 | 32.64 | 54.78 | 79.84 | |
| DeepSeek-R1-671B-0528Model=DeepSeek-R1-671B-0528, Category=Commercial Models2026.05 | 23.68 | 28.82 | 33.33 | |
| Qwen3-235B-A22B-InstructModel=Qwen3-235B-A22B-Instruct, Category=Commercial Models2026.05 | 22.21 | 24.22 | 26.67 | |
| Qwen2.5-7B GRPOModel=Qwen2.5-7B, Method=GRPO2026.05 | 19.32 | 35.02 | 58.1 | |
| Qwen2.5-3B EP-GRPOModel=Qwen2.5-3B, Method=EP-GRPO2026.05 | 11.2 | 21.23 | 38.2 | |
| Qwen2.5-7B BaseModel=Qwen2.5-7B, Method=Base2026.05 | 7.31 | 14.15 | 26.49 | |
| Qwen2.5-3B GRPO (More Rollouts)Model=Qwen2.5-3B, Method=GRPO, Rollouts=More2026.05 | 5.72 | 11.17 | 21.25 | |
| Qwen2.5-3B GRPO (Higher Temp)Model=Qwen2.5-3B, Method=GRPO, Temperature=Higher2026.05 | 4.11 | 8.09 | 15.65 | |
| Qwen2.5-3B BaseModel=Qwen2.5-3B, Method=Base2026.05 | 3.3 | 6.52 | 12.72 | |
| Qwen2.5-3B GRPOModel=Qwen2.5-3B, Method=GRPO2026.05 | 2.48 | 4.93 | 9.69 |