Process-level Evaluation on ProcessBench Omni
25.6F1 ScoreCPMI_Merge
Evaluation results
| Method | Links | |
|---|---|---|
| CPMI_MergeReward Type=CPMI_Merge2026.04 | 25.6 | |
| CPMIReward Type=CPMI2026.04 | 24.8 | |
| PAVReward Type=PAV2026.04 | 23.1 | |
| MSReward Type=MS2026.04 | 19.8 | |
| MCReward Type=MC2026.04 | 18.9 | |
| Rand_MergeReward Type=Rand_Merge2026.04 | 15.9 | |
| RLHFlowReward Type=RLHFlow2026.04 | 15.8 |