List your company

Mathematical Problem Solving on AIME 2024 (Pass@1 to Pass@1024)

30.3Pass@1CaSP
12.72417.28721.8526.413Oct 16, 2025

Evaluation results

MethodLinks
CaSPBackbone=Qwen2.5-Math-7B2025.1030.339.446.351.957.161.66671.176.581.285.6
GRPOBackbone=Qwen2.5-Math-7B2025.102733.339.24551.458.164.670.474.476.276.7
Base ModelBackbone=Qwen2.5-Math-7B2025.1013.421.329.636.742.949.155.260.866.17280.8