List your company

Mathematical Reasoning on AIME 25 (Accuracy, Average)

23.3AccuracyEEPO
9.46813.05916.6520.241Oct 7, 2025

Evaluation results

MethodLinks
EEPOConfiguration=Standard2025.1023.338.3
GRPOConfiguration=Higher Temp.2025.102035.2
GRPOConfiguration=Increased Ent.2025.102035.9
GRPOConfiguration=More rollouts.2025.1016.735.1
GRPOConfiguration=Standard2025.1016.634.7
Qwen3-8B-BaseConfiguration=Base Model2025.1013.329
GRPOConfiguration=DAPO Clip High.2025.101032.7