List your company

Mathematical Reasoning on HMMT 25 (pass@1, pass@64)

9Pass@1CurveRL
-0.1522.2244.66.976May 8, 2026May 10, 2026May 13, 2026May 15, 2026May 18, 2026May 20, 2026May 23, 2026

Evaluation results

MethodLinks
CurveRLBackbone=Qwen3-4B-Base2026.05932.4
MaxRLBackbone=Qwen3-4B-Base2026.057.927.8
GRPOBackbone=Qwen3-4B-Base2026.055.819
Zero-Shotn=N/A2026.054.1628.06
DPólyan=82026.053.6826.62
DPólyan=322026.053.6125.55
DPólyan=162026.053.626.52
DSTaRn=N/A2026.05228.84
CurveRLBackbone=Qwen3-1.7B-Base2026.051.719.3
DPólyan=42026.051.3926.02
MaxRLBackbone=Qwen3-1.7B-Base2026.051.318.4
BaseBackbone=Qwen3-4B-Base2026.05122.5
GRPOBackbone=Qwen3-1.7B-Base2026.050.510
BaseBackbone=Qwen3-1.7B-Base2026.050.210.6