ResearchBenchmarksFormal Theorem Proving on NuminaMath-LEAN (total)Follow51AccuracyAx-Prover3.1615.582840.42Oct 14, 2025Evaluation ResultsMethodMethodLinksAccuracyAx-Prover2025.1051Kiminaevaluation_protocol=pa...evaluation_protocol=pass@682025.1031DS-Prover2025.1028Sonnet2025.105