Mathematical Reasoning on AIME 24 (Pass@4, Pass@8, Pass@16)
42.6Pass@4 ScoreQwen2.5-7B EP-GRPO
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Qwen2.5-7B EP-GRPOModel=Qwen2.5-7B, Method=EP-GRPO2026.05 | 42.6 | 67.21 | 89.47 | |
| Qwen2.5-7B GRPOModel=Qwen2.5-7B, Method=GRPO2026.05 | 30.75 | 52.2 | 77.45 | |
| Qwen3-235B-A22B-InstructModel=Qwen3-235B-A22B-Instruct, Category=Commercial Models2026.05 | 28.32 | 30 | 33.33 | |
| DeepSeek-R1-671B-0528Model=DeepSeek-R1-671B-0528, Category=Commercial Models2026.05 | 28.07 | 32.14 | 33.33 | |
| Qwen2.5-3B EP-GRPOModel=Qwen2.5-3B, Method=EP-GRPO2026.05 | 22.82 | 40.56 | 64.99 | |
| Qwen2.5-3B GRPOModel=Qwen2.5-3B, Method=GRPO2026.05 | 17.16 | 31.49 | 53.37 | |
| Qwen2.5-3B GRPO (Higher Temp)Model=Qwen2.5-3B, Method=GRPO, Temperature=Higher2026.05 | 15.7 | 29.04 | 49.95 | |
| Qwen2.5-7B BaseModel=Qwen2.5-7B, Method=Base2026.05 | 14.96 | 27.79 | 48.15 | |
| Qwen2.5-3B GRPO (More Rollouts)Model=Qwen2.5-3B, Method=GRPO, Rollouts=More2026.05 | 11.2 | 21.23 | 38.2 | |
| Qwen2.5-3B BaseModel=Qwen2.5-3B, Method=Base2026.05 | 8.1 | 15.61 | 28.99 |