List your company

Mathematical Reasoning on AMC23 (Pass@1, Pass@32)

74.4Pass@1DynaMO
49.64856.07462.568.926Feb 22, 2026

Evaluation results

MethodLinks
DynaMOBackbone=Qwen2.5-Math-7B2026.0274.492.9
Clip-COVBackbone=Qwen2.5-Math-7B2026.0272.789.3
W-REINFORCEBackbone=Qwen2.5-Math-7B2026.0272.589.8
Fork TokensBackbone=Qwen2.5-Math-7B2026.027187.3
KL-COVBackbone=Qwen2.5-Math-7B2026.0270.688.5
GRPOBackbone=Qwen2.5-Math-7B2026.0268.390.8
Entropy AdvantagesBackbone=Qwen2.5-Math-7B2026.0267.985.2
Clip-HigherBackbone=Qwen2.5-Math-7B2026.0267.889.9
Entropy LossBackbone=Qwen2.5-Math-7B2026.026687
DynaMOBackbone=Qwen2.5-Math-1.5B2026.0263.691.9
W-REINFORCEBackbone=Qwen2.5-Math-1.5B2026.026385.7
Entropy AdvantagesBackbone=Qwen2.5-Math-1.5B2026.026286.4
Clip-COVBackbone=Qwen2.5-Math-1.5B2026.0259.589.7
GRPOBackbone=Qwen2.5-Math-1.5B2026.025690
KL-COVBackbone=Qwen2.5-Math-1.5B2026.0255.891.3
Entropy LossBackbone=Qwen2.5-Math-1.5B2026.0255.686.9
Fork TokensBackbone=Qwen2.5-Math-1.5B2026.0252.585.6
Clip-HigherBackbone=Qwen2.5-Math-1.5B2026.0250.689.9