ResearchDatasetsMulti-Task SuiteMulti-Task SuiteFollowBenchmarksTaskDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyTaskDatasetSOTA resultTrendResultsMultimodal ReasoningMulti-Task Suite VSP, Maze, TSP, Sudoku, Jigsaw, VisPuzzle88.5Average Score12Multi-task Performance EvaluationMulti-task Suite D1-D4 (test)94Avg F111