{"id":"tontaube-v0","nom":"Tontaube V0","editeur":"Tontaube","pays":"Allemagne","poids_ouverts":false,"poids_selon":null,"licence":null,"licence_selon":null,"liens":{"page":"https://tontaube.ai/blog/tontaube-v0-model-card"},"fiche_auto":true,"page":"https://quelleia.com/voix/modeles/tontaube-v0/","releve":"2026-10-19","en_bref":"31e sur 38 en synthèse vocale en anglais (TTS Arena).","presentation":"Premier modèle public de synthèse vocale de Tontaube, autorégressif, avec clonage de voix sans entraînement préalable. L'éditeur le décrit comme économe en calcul, plus rapide que le temps réel sur un processeur grand public et proche du temps réel sur smartphone.","sortie":"2026-04","langues":{"annoncees":null,"prises_en_charge_selon_open_tts":null,"evaluees":["en"]},"preuves":[{"appuie":["presentation","sortie"],"lu":"https://tontaube.ai/blog/tontaube-v0-model-card (« Model Card · April 2026 Tontaube V0 A compute-efficient text-to-speech model, built to match commercial voice-agent quality and run on-device. »), 17/10/2026"},{"appuie":["presentation"],"lu":"https://tontaube.ai/blog/tontaube-v0-model-card (« Tontaube V0 is an autoregressive text to speech model with zero-shot voice cloning. »), 17/10/2026"},{"appuie":["presentation"],"lu":"https://tontaube.ai/blog/tontaube-v0-model-card (« Trained for ~$300, runs faster than real-time on consumer CPU and approaches real-time on mobile. »), 17/10/2026"}],"classements":[{"id":"synthese/tts-arena","nom":"Synthèse vocale en anglais, TTS Arena","rang":31,"sur":38,"valeur":1481,"unite":"Elo","bas":1448,"haut":1514,"votes":519,"voisins":["kokoro-82m-v1-0","magpie-research-preview"],"source":{"nom":"TTS Arena V2","mention":"Source : TTS Arena V2, TTS-AGI","url":"https://huggingface.co/spaces/TTS-AGI/TTS-Arena-V2","releve":"2026-10-19"}}],"noms_dans_les_sources":{"tts_arena":"Tontaube V0"}}