ResearchBenchmarksRanking Prediction on MATH-plus GPT-4Follow0.71Kendall's TauPARS0.47080.53290.5950.6571Sep 25, 2025Evaluation ResultsMethodMethodLinksKendall's TauPARSRanking Approach=PairwiseRanking Approach=Pairwise2025.090.71Pointwise2025.090.54Listwise2025.090.48