ResearchTasksMulti-turn decision makingMulti-turn decision makingFollowBenchmarksDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyDatasetSOTA methodMetricTrendResultsLast updatedALFWorld (val-unseen)LoRA baseline94.1Success Rate18Mar 20, 2026tau2-Bench held-out (test)CURATEEVO41.9Average Score16Jul 8, 2026