ResearchDatasetsWeChat Mini-AppFollowBenchmarksTask NameDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyTask NameDataset NameSOTA ResultTrendResultsGUI Agent Task CompletionWeChat Mini-App Synthetic (train val)71.81VLM Eval Success Rate4GUI Agent Task CompletionWeChat Mini-App Real-World (test)42.28Success Rate (Human)4