ResearchTasksReasoning and Generative TasksReasoning and Generative TasksFollowBenchmarksDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyDatasetSOTA methodMetricTrendResultsLast updatedMMLU, CMMLU, GSM8k, XSum, and StrategyQA (test)GradMAP30.77MMLU Accuracy4Feb 27, 2026MMLU, CMMLU, GSM8k, XSum, and StrategyQA——MMLU0Feb 26, 2026