ResearchBenchmarksMathematical Reasoning on HMMT (Mean@16, Acc)Follow43.12Mean@16ME-ICPO-1.28810.24121.7733.299Mar 2, 2026Evaluation ResultsMethodMethodLinksMean@16AccuracyME-ICPOModel=Gemini-2.5-FlashModel=Gemini-2.5-Flash2026.0343.1243.14BaseModel=Gemini-2.5-FlashModel=Gemini-2.5-Flash2026.0314.7914.76BaseModel=Qwen2.5-Math-7BModel=Qwen2.5-Math-7B2026.031.040.67ME-ICPOModel=Qwen2.5-Math-7BModel=Qwen2.5-Math-7B2026.030.421.33