ResearchBenchmarksMathematical Reasoning on Olympiad OpenR1-Math Harder Subset (Accuracy)Follow68.1AccuracyRePO62.27663.78865.366.812Feb 11, 2026Evaluation ResultsMethodMethodLinksAccuracyRePOBackbone=Qwen, Paramet...Backbone=Qwen, Parameters=4B2026.0268.1Qwen-4BBackbone=Qwen, Paramet...Backbone=Qwen, Parameters=4B2026.0267.4LUFFYBackbone=Qwen, Paramet...Backbone=Qwen, Parameters=4B2026.0262.5