List your company

Mathematical Reasoning on MATH 500 (pass@1, pass@64)

85.1Pass@1MaxRL
55.9863.5471.178.66May 23, 2026

Evaluation results

MethodLinks
MaxRLBackbone=Qwen3-4B-Base2026.0585.196.6
CurveRLBackbone=Qwen3-4B-Base2026.0584.597.7
GRPOBackbone=Qwen3-4B-Base2026.0582.894.7
MaxRLBackbone=Qwen3-1.7B-Base2026.0572.194
CurveRLBackbone=Qwen3-1.7B-Base2026.0571.694.6
GRPOBackbone=Qwen3-1.7B-Base2026.0570.592.6
BaseBackbone=Qwen3-4B-Base2026.0567.595.7
BaseBackbone=Qwen3-1.7B-Base2026.0557.192.8