{"id":"vui-labs-luna-tts","nom":"Luna TTS","editeur":"VUI Labs","pays":"Hong Kong","poids_ouverts":false,"poids_selon":"Artificial Analysis","licence":null,"licence_selon":null,"liens":{"page":"https://vuilabs-ai.github.io/luna-tts","essayer":"https://www.vuilabs.ai/en"},"fiche_auto":true,"page":"https://quelleia.com/voix/modeles/vui-labs-luna-tts/","releve":"2026-10-19","en_bref":"13e sur 96 en synthèse vocale en anglais (Artificial Analysis), dans un peloton de 6 modèles que la marge d'erreur ne permet pas de séparer.","presentation":"Modèle de synthèse vocale non autorégressif à base de diffusion masquée, qui génère toute la grille de jetons audio en un nombre fixe d'étapes et permet nativement le clonage de voix sans entraînement et l'édition de la parole. Une phase d'affinage ajoute le contrôle des émotions et des vocalisations non verbales.","sortie":"2026-08-12","taille":"0,6 milliard de paramètres","langues":{"annoncees":{"annoncees":"4 langues (chinois, anglais, japonais, coréen)","nombre":4,"francais":false,"liste":["zh","en","ja","ko"]},"prises_en_charge_selon_open_tts":null,"evaluees":["en"]},"preuves":[{"appuie":["presentation"],"lu":"https://vuilabs-ai.github.io/luna-tts (« Luna-TTS generates the entire token grid in a fixed number of parallel refinement steps, with zero-shot voice cloning and speech editing arising natively as infilling. »), 17/10/2026"},{"appuie":["presentation"],"lu":"https://vuilabs-ai.github.io/luna-tts (« An annealed fine-tuning stage adds emotion and non-verbal vocalization control »), 17/10/2026"},{"appuie":["langues"],"lu":"https://vuilabs-ai.github.io/luna-tts (« 1 million hours of speech across Chinese, English, Japanese, and Korean »), 17/10/2026"},{"appuie":["langues"],"lu":"https://arxiv.org/html/2608.11593v1 (« Language coverage is also limited to four languages, substantially narrower than that of massively multilingual TTS systems. »), 17/10/2026"},{"appuie":["taille"],"lu":"https://vuilabs-ai.github.io/luna-tts (« On Seed-TTS-Eval, Luna-TTS leads all four metrics among the systems we compare, at 0.6B parameters »), 17/10/2026"},{"appuie":["sortie"],"lu":"https://arxiv.org/abs/2608.11593 (« [Submitted on 12 Aug 2026] »), 17/10/2026"},{"appuie":["sortie"],"lu":"https://arxiv.org/html/2608.11593v1 (« Luna-TTS Family Technical Report VUI Labs Research »), 17/10/2026"},{"appuie":["presentation"],"lu":"https://www.vuilabs.ai/ (« Luna-TTS $15 / 1,000,000 characters High-Quality, Ultra-Natural Speech High-Fidelity Voice Cloning »), 17/10/2026"}],"classements":[{"id":"synthese/artificial-analysis","nom":"Synthèse vocale en anglais, Artificial Analysis","rang":13,"sur":96,"valeur":1216.96,"unite":"Elo","bas":1204.96,"haut":1228.96,"votes":3382,"chiffres":{"prix_par_million_caracteres":15,"date_sortie":"2026-06-18"},"voisins":["inworld-realtime-tts-2-flash","stepaudio-2-5-tts-aug-2026"],"source":{"nom":"Artificial Analysis","mention":"Source : Artificial Analysis","url":"https://artificialanalysis.ai/text-to-speech/model-families/luna-tts","releve":"2026-10-19"}}],"noms_dans_les_sources":{"artificial_analysis":"Luna TTS"}}