List your company

Mathematical Reasoning on MATH (test) (Pass@k Suite)

77.2Pass@1AGPO
62.53666.34370.1573.957May 7, 2026

Evaluation results

MethodLinks
AGPOBackbone=Qwen2.5-Math-7B2026.0577.283.587.990.792.794.395.696.496.6
W-REINFORCEBackbone=Qwen2.5-Math-7B2026.0576.382.486.990.292.393.895.496.296.6
GRPOBackbone=Qwen2.5-Math-7B2026.0576.281.585.2889091.7939494.7
PPOBackbone=Qwen2.5-Math-7B2026.0575.881.585.588.490.692.293.494.394.8
REINFORCEBackbone=Qwen2.5-Math-7B2026.0574.178.781.984.486.888.289.69192.3
Base ModelBackbone=Qwen2.5-Math-7B2026.0563.176.283.588.191.593.695.396.296.7