List your company

Math Reasoning on AIME 25 (pass@1, cons@32)

35.72Pass@1XRPO
25.652828.266430.8833.4936Oct 8, 2025

Evaluation results

MethodLinks
XRPOBackbone=Qwen3-1.7B2025.1035.7250
GSPOBackbone=Qwen3-1.7B2025.1033.2346.67
DAPOBackbone=Qwen3-1.7B2025.1032.533.33
TPO-SBackbone=Qwen3-1.7B2025.1026.0433.33