ResearchTasksLong-form audio-video understandingFollowBenchmarksDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyDataset NameSOTA methodMetricTrendResultsLast UpdatedOmniVideoBench up to 30minMiniCPM-o 2.626.2Latency (10, 30] min9Jun 24, 2026WorldSense up to 10min (Avg.)AVOC52.6Avg Performance (WorldSense)8Jun 24, 2026LVOmniBench up to 90minAVOC41.3Performance Score (Medium Tier)7Jun 24, 2026