ResearchBenchmarksReasoning on MATH (avg.@8)Follow87.48avg.@8 ScoreMT62.436868.938475.4481.9416Apr 22, 2026Evaluation ResultsMethodMethodLinksavg.@8 ScoreMTModel Family=Qwen 2.5,...Model Family=Qwen 2.5, Model Size=7B2026.0487.48MTModel Family=LLaMA 3,...Model Family=LLaMA 3, Model Size=8B2026.0477.78HPDModel Family=Qwen 2.5,...Model Family=Qwen 2.5, Student Model=3B2026.0476.3HPDModel Family=LLaMA 3,...Model Family=LLaMA 3, Student Model=3B2026.0469.25HPDModel Family=Qwen 2.5,...Model Family=Qwen 2.5, Student Model=1.5B2026.0463.4