ResearchTasksPairwise comparison evaluationPairwise comparison evaluationFollowBenchmarksDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyDatasetSOTA methodMetricTrendResultsLast updatedArena Human Preference DataCollabEval51.5Accuracy5Mar 4, 2026Chatbot ArenaCollabEval60.2Accuracy5Mar 4, 2026