List your company

Mathematical Reasoning on Math Domain

66.45Avg AccuracyDVPO
58.07860.251562.42564.5985Dec 3, 2025

Evaluation results

MethodLinks
DVPOTraining Domain=Math Domain2025.1266.45
Dr.GRPOTraining Domain=Math Domain2025.1261.64
Reinforce++Training Domain=Math Domain2025.1261.35
Robust BellmanTraining Domain=Math Domain2025.1260.56
GRPOTraining Domain=Math Domain2025.1259.26
BaseTraining Domain=Math Domain2025.1258.4
PPOTraining Domain=Math Domain2025.1258.4