ResearchTasksMulti-modal Task EvaluationFollowBenchmarksDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyDataset NameSOTA methodMetricTrendResultsLast UpdatedMM-MergeBench Seen TasksIndividual83.74SciQA Score12Mar 9, 2026MM-MergeBench Unseen TasksRobustMerge79.3AVQA Score11Mar 9, 2026