List your company

Mathematical Reasoning on GSM8K (Accuracy Percentiles & Token Costs)

93.07Mean Accuracy (±STD)ACTIVE
92.23892.45492.6792.886May 7, 2026

Evaluation results

MethodLinks
ACTIVESetup=ACTIVE2026.0593.0792.8292.772.10.183.54
RANDOMSetup=RANDOM2026.0592.7892.4992.06—0.173.56
NO TRAINSetup=NO TRAIN2026.0592.2791.7391.61——4.5