Mathematical Reasoning on ProofNet
97.2AccuracyPhi-4 pass@N (Upper Bound)
Evaluation Results
| Method | Links | |
|---|---|---|
| Phi-4 pass@N (Upper Bound)# Sample=-2025.11 | 97.2 | |
| Skywork-PRM-1.5B# Sample=Unk2025.11 | 95.7 | |
| Math-Shepherd-PRM-7B# Sample=440K2025.11 | 95.2 | |
| Qwen2.5-Math-7B-PRM# Sample=860K2025.11 | 94.4 | |
| Universal-PRM-Qwen2.5-Math-7B# Sample=690K2025.11 | 93.5 | |
| ReProbe, Attn+Logit, Qwen3-8B-anno# Sample=32K2025.11 | 93.1 | |
| Qwen2.5-Math-7B-PRM800k# Sample=263K2025.11 | 92.4 | |
| RLHFlow-PRM-Mistral-Data# Sample=273K2025.11 | 90.3 | |
| RLHFlow-PRM-Deepseek-Data# Sample=253K2025.11 | 89.6 | |
| H4-Qwen2.5-PRM-1.5B-0.2# Sample=369K2025.11 | 89.2 | |
| Phi-4 pass@1 (Lower Bound)# Sample=-2025.11 | 86.8 |