ResearchTasksMulti-task Question AnsweringMulti-task Question AnsweringFollowBenchmarksDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyDatasetSOTA methodMetricTrendResultsLast updatedARC, OBQA, and Riddle Combined (test)REALM62.69Accuracy54Apr 21, 2026MMLUMistral Small 24B Inst 250178.7Accuracy46Jun 18, 2026GraphBench Novel 2,010 queries (test)CogitoRAG61.48Fact Retrieval (FR)8Feb 26, 2026