SimpleOPD: Simple Tokenizer-Agnostic On-Policy Distillation for Long-Context Reasoning

Benchmark Model Rank Results
mathematical-reasoning-on-aime-2025Intern-S2-DS-OPD#1Accuracy: 97.5
mathematical-reasoning-on-amo-benchQwen3.5-35B-A3B-OPD#1Accuracy: 61.25
science-question-answering-on-frontierscience-researchIntern-S2-OPD#5Accuracy: 5.0
science-question-answering-on-frontierscience-olympiadIntern-S2-OPD#7Accuracy: 60.9