ResearchTasksVideo ConversationFollowBenchmarksDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyDataset NameSOTA methodMetricTrendResultsLast UpdatedVideo-ChatGPT benchmark ActivityNet-200 (test)MiniGPT4-video3.08Correctness10Feb 26, 2026VCGBench-Diverse (test)VideoGPT+2.46CI Score6Feb 26, 2026VideoChatGPTGPT-4V4.06Aggregate Score4Feb 26, 2026