Mathematical Reasoning on MATH-500 (Pass@1 and Avg. Pass@1)
80.2Pass@1DS
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| DSModel=Qwen2.5-Math-7B, Avg. Rollouts=2.5x2026.02 | 80.2 | 39.4 | |
| MoPPSModel=Qwen2.5-Math-7B, Avg. Rollouts=1.0x2026.02 | 80.2 | 38.6 | |
| PACED-RLModel=Qwen2.5-Math-7B, Avg. Rollouts=1.0x2026.02 | 80 | 40.1 | |
| LILOModel=Qwen2.5-Math-7B, Avg. Rollouts=4.0x2026.02 | 78.4 | 40.1 | |
| GRPOModel=Qwen2.5-Math-7B, Avg. Rollouts=1.0x2026.02 | 77.6 | 37.2 | |
| FlowRLModel=Qwen2.5-Math-7B, Avg. Rollouts=1.0x2026.02 | 76.4 | 36.5 | |
| LILOModel=Qwen2.5-Math-1.5B, Avg. Rollouts=4.0x2026.02 | 73.2 | 31.2 | |
| PACED-RLModel=Qwen2.5-Math-1.5B, Avg. Rollouts=1.0x2026.02 | 73.2 | 32.1 | |
| DSModel=Qwen2.5-Math-1.5B, Avg. Rollouts=2.0x2026.02 | 72.8 | 31.2 | |
| MoPPSModel=Qwen2.5-Math-1.5B, Avg. Rollouts=1.0x2026.02 | 71.6 | 30.4 | |
| GRPOModel=Qwen2.5-Math-1.5B, Avg. Rollouts=1.0x2026.02 | 70.6 | 29.1 | |
| FlowRLModel=Qwen2.5-Math-1.5B, Avg. Rollouts=1.0x2026.02 | 67.6 | 27.1 |