ResearchBenchmarksHuman Evaluation on Education subsetFollow85Win RateMENTOR57.9664.987279.02Nov 10, 2025Evaluation ResultsMethodMethodLinksWin RateTie RateLoss RateMENTORBackbone=DeepSeek-R1Backbone=DeepSeek-R12025.1185132MENTORBackbone=GPT-4oBackbone=GPT-4o2025.1178220MENTORBackbone=Qwen3-235BBackbone=Qwen3-235B2025.1159356