List your company

Mathematical Reasoning on ProcessBench MATH 1.0 (test)

88.4AccuracyGreedy EDU PRM
54.91263.60672.380.994Mar 28, 2025

Evaluation results

MethodLinks
Greedy EDU PRMModel Size=72B2025.0388.488.290.486.2
Qwen2.5-Math-PRMModel Size=72B2025.0387.887.291.883
Qwen2.5-Math-PRMModel Size=7B2025.0382.48283.980.2
Omega PRMModel Size=72B2025.0379.876.392.365
Math-Shepherd PRMModel Size=72B2025.0377.880.572.790.2
Sample EDU PRMModel Size=72B2025.0376.479.570.990.6
Math-Shepherd PRMModel Size=7B2025.0362.965.961.571
Sample EDU PRMModel Size=7B2025.0359.268.955.989.8
Omega PRMModel Size=7B2025.035829.591.717.6
Greedy EDU PRMModel Size=7B2025.0356.222.995.613