ResearchTasksOffline GUI Agent EvaluationOffline GUI Agent EvaluationFollowBenchmarksDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyDatasetSOTA methodMetricTrendResultsLast updatedCAGUI (Full)Mimo-VL-7B + WildGUI90.3Type Accuracy9May 15, 2026AndroidControl HighUI-TARS-7B83.7Type Accuracy9May 15, 2026AndroidControl LowUI-TARS-7B98Action Type Accuracy9May 15, 2026