List your company

Mathematical Reasoning on MATH 500 (Pass@1, Pass@16, Major@16)

86.8Pass@1GSPO
59.34466.47273.680.728May 18, 2026

Evaluation results

MethodLinks
GSPOBackbone=Qwen3-4B2026.0586.893.290
CorR-POBackbone=Qwen3-4B2026.0586.592.489.8
DAPOBackbone=Qwen3-4B2026.0586.293.290.2
GRPOBackbone=Qwen3-4B2026.0583.49388.2
Dr.GRPOBackbone=Qwen3-4B2026.05799286.4
BaseBackbone=Qwen3-4B2026.0560.490.879.8