List your company

Math Reasoning on GSM8K (Pass@K)

83.2Pass@1 AccuracyMIG
77.58479.04280.581.958Feb 1, 2026

Evaluation results

MethodLinks
MIGEvaluation Mode=In-Domain, Training Strategy=Task-Specific Training2026.0283.295.62.2
GRPOEvaluation Mode=In-Domain, Training Strategy=Task-Specific Training2026.028195.6—
Base ModelEvaluation Mode=In-Domain, Training Strategy=Task-Specific Training2026.0277.893.2—