ResearchDatasetsCalendarFollowBenchmarksTask NameDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyTask NameDataset NameSOTA ResultTrendResultsStep-level correctness assessmentCalendar (test)0.789PR-AUC22Step-level reasoning verificationCalendar69.8PR-AUC19PlanningCalendar (test)48Accuracy6Skill evolutionCalendar merge77.7Best@201