ResearchTasksGUI reasoningFollowBenchmarksDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyDataset NameSOTA methodMetricTrendResultsLast UpdatedAndroidControl LowUI-TARS-7B91.8SR24Apr 16, 2026Aggregate (GUI-Act-Web, OmniAct-Web, AndroidControl)CAPO74.6Overall Accuracy4Feb 27, 2026AndroidControl HighCAPO65.91Type4Feb 27, 2026OmniAct-WebCAPO87.24Type Success Rate4Feb 27, 2026GUI-Act-WebCAPO87.73Type Success Rate4Feb 27, 2026