ResearchTasksCorrelation analysis of reasoning metrics with ground-truth accuracyFollowBenchmarksDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyDataset NameSOTA methodMetricTrendResultsLast Updated39 standard benchmarks AIME GSM8K ARC MMLU MMLU-PRO GPQA SuperGPQATRACE0.741Pearson r4May 29, 2026