List your company

Mathematical Reasoning on OlympiadBench (pass@3)

64.76Pass@3HDPO
37.844844.832451.8258.8076Jun 2, 2026

Evaluation results

MethodLinks
HDPOBackbone=Qwen2.5-Math-7B, Number of samples=162026.0664.76
HDPOBackbone=Qwen3-4B, Number of samples=162026.0662.37
Critique-GRPOBackbone=8B, Number of samples=162026.0661.46
HDPOBackbone=DS-R1-Llama-8B, Number of samples=162026.0656.19
LUFFYBackbone=Qwen-Math-7B, Number of samples=162026.0656.09
GRPOBackbone=Qwen3-4B, Number of samples=162026.0655.5
GRPOBackbone=DS-R1-Llama-8B, Number of samples=162026.0654.74
GRPOBackbone=Qwen2.5-Math-7B, Number of samples=162026.0653.68
OatBackbone=Qwen2.5-Math-7B, Number of samples=162026.0650.82
SimpleRLBackbone=Qwen-2.5-7B, Number of samples=162026.0649.84
EDGE-GRPOBackbone=Qwen-7B, Number of samples=162026.0648.79
Base ModelBackbone=Qwen2.5-Math-7B, Number of samples=162026.0643.73
Base ModelBackbone=Qwen3-4B, Number of samples=162026.0640.11
Base ModelBackbone=DS-R1-Llama-8B, Number of samples=162026.0638.88