ResearchTasksTarget Viewpoint ReproductionTarget Viewpoint ReproductionFollowBenchmarksDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyDatasetSOTA methodMetricTrendResultsLast updatedTVRBench 500-task (test)Human93Overall Success Rate16Jun 2, 2026TVRBench (test)Multi-turn GRPO51.4Overall Performance8Jun 2, 2026