ResearchTasksAgent Performance EvaluationFollowBenchmarksDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyDataset NameSOTA methodMetricTrendResultsLast UpdatedHarness-BenchFairyClaw74.6Combined Score12May 20, 2026MCP-BenchReAct46.8Task Fulfillment7May 12, 2026