ResearchTasksText-to-Speech SynthesisText-to-Speech SynthesisFollowBenchmarksDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyDatasetSOTA methodMetricTrendResultsLast updatedLibriTTS (CLEAN), LibriVox (NOISY), YouTube (WILD), and My Science Tutor (KIDS) (test)Ground Truth3.7MOS21Mar 4, 2026LJSpeechFNH-TTS4.48MOS11May 29, 2026VCTKFNH-TTS4.63MOS9May 29, 202610-second speech segmentsMELLE-R41.4Inference Time (s)8Feb 26, 2026Internal (test)Recording4.16MOS4Jun 16, 2026SEED-TTS en | zh | avg.UniAudio-Token0.792SIM (en)2Jun 1, 2026
LibriTTS (CLEAN), LibriVox (NOISY), YouTube (WILD), and My Science Tutor (KIDS) (test)Ground Truth3.7MOS21Mar 4, 2026