Mathematical Reasoning on Aggregate (MATH500, AMC23, AIME24, AIME25, BBEH)
44.8Average ScoreDenoiseRL-DAPO
Evaluation Results
| Method | Links | |
|---|---|---|
| DenoiseRL-DAPOBase Model=Qwen3-8B-Base2026.05 | 44.8 | |
| DenoiseRL-GRPOBase Model=Qwen3-8B-Base2026.05 | 43.3 | |
| GRPOBase Model=Qwen3-8B-Base2026.05 | 43 | |
| DAPOBase Model=Qwen3-8B-Base2026.05 | 42.8 | |
| DenoiseRL-GRPOBase Model=Qwen3-4B-Base2026.05 | 42 | |
| DenoiseRL-DAPOBase Model=Qwen3-4B-Base2026.05 | 41.5 | |
| DAPOBase Model=Qwen3-4B-Base2026.05 | 39.8 | |
| GRPOBase Model=Qwen3-4B-Base2026.05 | 39.6 | |
| BaseBase Model=Qwen3-8B-Base2026.05 | 29.3 | |
| BaseBase Model=Qwen3-4B-Base2026.05 | 26.6 |