List your company

Process Reward Modeling on PRM800K (test)

95.5Pos-AccReasonEval-7B*
50.5262.197573.87585.5525Jun 8, 2026

Evaluation results

MethodLinks
ReasonEval-7B*Number of Parameters=7B2026.0695.595.521.221.260
Qwen-PRM-7BNumber of Parameters=7B2026.0695.3695.3630.6630.6665.5
ThinkPRM-7BNumber of Parameters=7B2026.0683.2983.2950.8950.8964.3
GPT-4o*2026.0682.982.958.258.266.8
GenPRM-7BNumber of Parameters=7B2026.0652.2552.2573.9273.9250.5