Reasoning on ace
59.7Pass RateGRPO
Evaluation Results
| Method | Links | |
|---|---|---|
| GRPOModel=OLMo-3-7B-Instruct2026.05 | 59.7 | |
| SRPOModel=OLMo-3-7B-Instruct2026.05 | 59.6 | |
| RRPOModel=OLMo-3-7B-Instruct2026.05 | 59 | |
| Cr-GRPOModel=OLMo-3-7B-Instruct2026.05 | 50.5 | |
| SPO-TreeModel=OLMo-3-7B-Instruct2026.05 | 50.2 | |
| BaseModel=OLMo-3-7B-Instruct2026.05 | 48.4 | |
| SRPOModel=Qwen2.5-14B-Instruct2026.05 | 46.2 | |
| RRPOModel=Qwen2.5-14B-Instruct2026.05 | 45.7 | |
| GRPOModel=Qwen2.5-14B-Instruct2026.05 | 45.2 | |
| BaseModel=Qwen2.5-14B-Instruct2026.05 | 44.2 | |
| SCoReModel=Qwen2.5-14B-Instruct2026.05 | 43.7 | |
| SPO-TreeModel=Qwen2.5-14B-Instruct2026.05 | 43.1 | |
| Cr-GRPOModel=Qwen2.5-14B-Instruct2026.05 | 42.5 | |
| SCoReModel=OLMo-3-7B-Instruct2026.05 | 33.9 |