List your company

Mathematical Reasoning on AIME 24 (Accuracy, Average)

20AccuracyGRPO
9.612.31517.7Oct 7, 2025

Evaluation results

MethodLinks
GRPOConfiguration=Standard2025.102034.7
EEPOConfiguration=Standard2025.102038.3
GRPOConfiguration=Higher Temp.2025.1016.735.2
GRPOConfiguration=Increased Ent.2025.1016.735.9
GRPOConfiguration=DAPO Clip High.2025.1016.732.7
GRPOConfiguration=More rollouts.2025.1016.735.1
Qwen3-8B-BaseConfiguration=Base Model2025.101029