ResearchTasksReasoning FidelityReasoning FidelityFollowBenchmarksDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyDatasetSOTA methodMetricTrendResultsLast updatedAIME 2025 (ref6)CASK0.917Top-1 Accuracy6Apr 14, 2026AIME 2024 (ref6)CASK91.7Top-1 Acc6Apr 14, 2026