ResearchBenchmarksMathematical Reasoning on AIME 24 (Score)Follow3.33ScoreTD-MNPO-0.13320.76591.6652.5641Sep 27, 2025Evaluation ResultsMethodMethodLinksScoreTD-MNPO2025.093.33HT-MNPOReward Model=ArmoRM-Ll...Reward Model=ArmoRM-Llama32025.093.33SFT Model2025.090DPO2025.090SimPO2025.090SPPO2025.090INPO2025.090HT-MNPOReward Model=Skywork-R...Reward Model=Skywork-Reward-V22025.090HT-MNPOReward Model=Athene-RM-8BReward Model=Athene-RM-8B2025.090