ResearchDatasetsReal-phone Human EvaluationReal-phone Human EvaluationFollowBenchmarksTaskDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyTaskDatasetSOTA resultTrendResultsAgentic Phone Use150-task Real-phone Human Evaluation Single-App62Task Success Rate14Agentic Phone Use150-task Real-phone Human Evaluation Cross-App48Task Success Rate7