ResearchBenchmarksMathematical Reasoning on AIME 2025 (Avg@32 accuracy)Follow19.16Avg@32 AccuracyESPO14.937616.033817.1318.2262Nov 29, 2025Evaluation ResultsMethodMethodLinksAvg@32 AccuracyESPO2025.1119.16GSPO2025.1115.6DAPO2025.1115.1