ResearchTasksMulti-discipline Video UnderstandingFollowBenchmarksDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyDataset NameSOTA methodMetricTrendResultsLast UpdatedSeedBench videoLLaVA-OneVision-72B62.1Accuracy11Feb 26, 2026MLVU OverallFlashVID53.8MLVU Score10May 12, 2026MMVUGPT-4o75.4Accuracy7May 26, 2026VideoMMEGPT-4o71.9Accuracy5Feb 26, 2026