ResearchTasksAggregated PerformanceAggregated PerformanceFollowBenchmarksDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyDatasetSOTA methodMetricTrendResultsLast updatedAverage 10 TasksFlashVLM100.6Average Accuracy19Feb 27, 2026Downstream Average AllNexus40.3Accuracy4Apr 13, 2026