{"id":"cartesia-sonic-3","nom":"Sonic 3","editeur":"Cartesia","pays":"États-Unis","poids_ouverts":false,"poids_selon":"Artificial Analysis","licence":null,"licence_selon":null,"liens":{"page":"https://docs.cartesia.ai/build-with-cartesia/tts-models/older-models","essayer":"https://docs.cartesia.ai/api-reference/tts/bytes"},"fiche_auto":true,"page":"https://quelleia.com/voix/modeles/cartesia-sonic-3/","releve":"2026-10-19","en_bref":"20e sur 31 en synthèse vocale en français, dans un peloton de 5 modèles que la marge d'erreur ne permet pas de séparer ; 59e sur 96 en synthèse vocale en anglais (Artificial Analysis), dans un groupe de 13 modèles trop proches pour être départagés.","presentation":"Modèle de synthèse vocale temps réel de Cartesia, lancé en octobre 2025 avec 42 langues. Il ajoute des réglages de volume, de vitesse et d'émotion, et peut produire des rires dans la conversation.","sortie":"2025-10-27","langues":{"annoncees":{"annoncees":"42 langues, dont le français","nombre":42,"francais":true},"prises_en_charge_selon_open_tts":null,"evaluees":["fr","en"]},"preuves":[{"appuie":["langues","presentation","sortie"],"lu":"https://docs.cartesia.ai/changelog/2025 (« sonic-3-2025-10-27 stable snapshot released; 42 languages; volume, speed, and emotion controls. Real-time conversation with emotion and laughter; ~190ms median latency. »), 17/10/2026"},{"appuie":["langues"],"lu":"https://docs.cartesia.ai/build-with-cartesia/tts-models/older-models (« sonic-3-2026-01-12 January 12, 2026 en, de, es, fr, ja, pt, zh, hi, ko »), 17/10/2026"}],"classements":[{"id":"synthese/francais","nom":"Synthèse vocale en français","rang":20,"sur":31,"valeur":1097.6,"unite":"Elo","bas":1076.6,"haut":1118.6,"votes":1161,"chiffres":{"prix_par_million_caracteres":49,"date_sortie":"2025-10-29"},"voisins":["minimax-speech-2-8-hd","minimax-speech-2-8-turbo"],"source":{"nom":"Artificial Analysis","mention":"Source : Artificial Analysis","url":"https://artificialanalysis.ai/text-to-speech/model-families/cartesia","releve":"2026-10-19"}},{"id":"synthese/artificial-analysis","nom":"Synthèse vocale en anglais, Artificial Analysis","rang":59,"sur":96,"valeur":1057.64,"unite":"Elo","bas":1045.64,"haut":1069.64,"votes":3539,"chiffres":{"prix_par_million_caracteres":49,"date_sortie":"2025-10-29"},"voisins":["google-gemini-2-5-flash-tts-dec-2025","rime-coda"],"source":{"nom":"Artificial Analysis","mention":"Source : Artificial Analysis","url":"https://artificialanalysis.ai/text-to-speech/model-families/cartesia","releve":"2026-10-19"}}],"noms_dans_les_sources":{"artificial_analysis":"Sonic 3"}}