List your company

Mathematical Reasoning on OlympiadBench (Pass@1 accuracy)

70.7Pass@1 AccuracyPADD
30.1440.6751.261.73Jun 9, 2026

Evaluation results

MethodLinks
PADDModel Family=Qwen Family2026.0670.7
RSPOModel Family=Qwen Family2026.0666.2
Teacher (GRPO)Model Family=Qwen Family2026.0663.8
GSPOModel Family=Qwen Family2026.0663.6
BaseModel Family=Qwen Family2026.0662.2
Online KDModel Family=Qwen Family2026.0659.6
Vanilla-GRPOModel Family=Qwen Family2026.0658.2
Teacher (GRPO)Model Family=DeepSeek Family2026.0649.7
PADDModel Family=DeepSeek Family2026.0649.7
RSPOModel Family=DeepSeek Family2026.0649.6
GSPOModel Family=DeepSeek Family2026.0647.9
Online KDModel Family=DeepSeek Family2026.0643.7
Vanilla-GRPOModel Family=DeepSeek Family2026.0641.8
Dense-GRPOModel Family=Qwen Family2026.0640.9
Dense-GRPOModel Family=DeepSeek Family2026.0635.9
BaseModel Family=DeepSeek Family2026.0631.7