ResearchTasksMulti-modal to Audio Generation LatencyMulti-modal to Audio Generation LatencyFollowBenchmarksDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyDatasetSOTA methodMetricTrendResultsLast updatedServeGen 16-GPU setup (real-world workload)Cornserve21.3P50 Latency8Feb 26, 2026