Mathematical Reasoning on AIME25 (pass@1 accuracy)
76Pass@1 AccuracySRGen
Evaluation results
| Method | Links | |
|---|---|---|
| SRGenModel=Qwen3-32B2025.10 | 76 | |
| CoTModel=Qwen3-32B2025.10 | 70.7 | |
| Self-RefineModel=Qwen3-32B2025.10 | 69.3 | |
| SRGenModel=DS-R1-Qwen-7B2025.10 | 42.7 | |
| Self-RefineModel=DS-R1-Qwen-7B2025.10 | 38 | |
| SRGenModel=DS-R1-Llama-8B2025.10 | 36 | |
| CoTModel=DS-R1-Qwen-7B2025.10 | 35.3 | |
| Self-RefineModel=DS-R1-Llama-8B2025.10 | 32 | |
| CoTModel=DS-R1-Llama-8B2025.10 | 30.7 | |
| SRGenModel=Qwen2.5-Math-7B2025.10 | 9.3 | |
| CoTModel=Qwen2.5-Math-7B2025.10 | 6 | |
| Self-RefineModel=Qwen2.5-Math-7B2025.10 | 5.7 |