Mathematical Reasoning on HMMT (Pass Rate %)
15.6Pass RateSRPO
Evaluation Results
| Method | Links | |
|---|---|---|
| SRPOModel=OLMo-3-7B-Instruct2026.05 | 15.6 | |
| SPO-TreeModel=OLMo-3-7B-Instruct2026.05 | 12.2 | |
| GRPOModel=OLMo-3-7B-Instruct2026.05 | 11.1 | |
| BaseModel=OLMo-3-7B-Instruct2026.05 | 10 | |
| Cr-GRPOModel=OLMo-3-7B-Instruct2026.05 | 10 | |
| RRPOModel=OLMo-3-7B-Instruct2026.05 | 10 | |
| Cr-GRPOModel=Qwen2.5-14B-Instruct2026.05 | 7.8 | |
| RRPOModel=Qwen2.5-14B-Instruct2026.05 | 7.8 | |
| BaseModel=Qwen2.5-14B-Instruct2026.05 | 6.7 | |
| SRPOModel=Qwen2.5-14B-Instruct2026.05 | 6.7 | |
| GRPOModel=Qwen2.5-14B-Instruct2026.05 | 5.6 | |
| SCoReModel=Qwen2.5-14B-Instruct2026.05 | 5.6 | |
| SPO-TreeModel=Qwen2.5-14B-Instruct2026.05 | 4.4 | |
| SCoReModel=OLMo-3-7B-Instruct2026.05 | 0 |