List your company

Mathematical Reasoning on GSM8K, MATH500, OlympiadBench, AIME (test)

71.4Pass@1DRPO
52.6857.5462.467.26Oct 6, 2025

Evaluation results

MethodLinks
DRPOModel Size=7B2025.1071.41,5020.249
RLOO-LPModel Size=7B2025.1069.22,649-0.033
DRPOModel Size=8B2025.10691,7580.297
ALPModel Size=7B2025.1067.92,170-0.134
ALPModel Size=8B2025.1063.72,358-0.01
DRPOModel Size=1.5B2025.1062.41,5270.178
RLOO-LPModel Size=8B2025.1061.92,2490.251
ALPModel Size=1.5B2025.1060.63,494-0.387
HAPOModel Size=7B2025.1059.61,717-1.08
HAPOModel Size=8B2025.1059.51,981-0.366
RLOO-LPModel Size=1.5B2025.1056.72,531-0.129
HAPOModel Size=1.5B2025.1053.41,791-0.519