Process Reward Modeling on PRM800K (test)
95.5Pos-AccReasonEval-7B*
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| ReasonEval-7B*Number of Parameters=7B2026.06 | 95.5 | 95.5 | 21.2 | 21.2 | 60 | |
| Qwen-PRM-7BNumber of Parameters=7B2026.06 | 95.36 | 95.36 | 30.66 | 30.66 | 65.5 | |
| ThinkPRM-7BNumber of Parameters=7B2026.06 | 83.29 | 83.29 | 50.89 | 50.89 | 64.3 | |
| GPT-4o*2026.06 | 82.9 | 82.9 | 58.2 | 58.2 | 66.8 | |
| GenPRM-7BNumber of Parameters=7B2026.06 | 52.25 | 52.25 | 73.92 | 73.92 | 50.5 |