{"id":"openaudio-s1","nom":"OpenAudio S1","editeur":"Fish Audio","pays":"États-Unis","poids_ouverts":false,"poids_selon":null,"licence":null,"licence_selon":null,"liens":{"page":"https://fish.audio/blog/introducing-s1/","essayer":"https://fish.audio/app/text-to-speech/"},"fiche_auto":false,"page":"https://quelleia.com/voix/modeles/openaudio-s1/","releve":"2026-10-19","en_bref":"28e sur 31 en synthèse vocale en français, ex æquo avec 2 autres modèles dont les marges d'erreur se recouvrent ; 47e sur 96 en synthèse vocale en anglais (Artificial Analysis), dans un peloton de 10 modèles que la marge d'erreur ne permet pas de séparer ; 24e sur 38 en synthèse vocale en anglais (TTS Arena).","presentation":"Modèle de synthèse vocale de 4 milliards de paramètres présenté sous la marque OpenAudio de Fish Audio, entraîné sur plus de 2 millions d'heures d'audio avec apprentissage par renforcement (GRPO). Il se pilote par des marqueurs d'émotion, de ton et d'effets sonores, et existe en version réduite S1-mini à poids ouverts.","sortie":"2025-06-03","taille":"4 milliards de paramètres","langues":{"annoncees":{"annoncees":"13 langues, dont le français","nombre":13,"francais":true,"liste":["en","zh","ja","de","fr","es","ko","ar","ru","nl","it","pl","pt"]},"prises_en_charge_selon_open_tts":null,"evaluees":["fr","en"]},"preuves":[{"appuie":["presentation","sortie"],"lu":"https://openaudio.com/blogs/s1 (« 2025-06-03 · Research Introducing S1 Our cutting-edge text-to-speech model that performs like voice actors Try it at Fish Audio We are incredibly excited to unveil OpenAudio S1 »), 17/10/2026"},{"appuie":["presentation","taille"],"lu":"https://openaudio.com/blogs/s1 (« scaling our model to 4B parameters, incorporating our in-house reward modeling, and applying RLHF (GRPO) training »), 17/10/2026"},{"appuie":["langues"],"lu":"https://fish.audio/blog/introducing-s1/ (« OpenAudio S1 is designed for global reach. It supports a wide range of languages, including: English, Chinese, Japanese, German, French, Spanish, Korean, Arabic, Russian, Dutch, Italian, Polish, Portuguese »), 17/10/2026"},{"appuie":["langues"],"lu":"https://docs.fish.audio/developer-guide/models-pricing/models-overview.md (« S1 supports text-to-speech generation in 13 languages with full emotional expression capabilities. »), 17/10/2026"},{"appuie":["presentation"],"lu":"https://huggingface.co/fishaudio/s1-mini (« S1 (4B, proprietary): The full-sized model. S1-mini (0.5B): A distilled version of S1. »), 17/10/2026"}],"classements":[{"id":"synthese/francais","nom":"Synthèse vocale en français","rang":28,"sur":31,"valeur":1000,"unite":"Elo","bas":1000,"haut":1000,"votes":1188,"chiffres":{"prix_par_million_caracteres":15,"date_sortie":"2025-06-03"},"voisins":["coqui-xtts-v2","fish-audio-s1"],"source":{"nom":"Artificial Analysis","mention":"Source : Artificial Analysis","url":"https://artificialanalysis.ai/text-to-speech/model-families/fishaudio","releve":"2026-10-19"}},{"id":"synthese/artificial-analysis","nom":"Synthèse vocale en anglais, Artificial Analysis","rang":47,"sur":96,"valeur":1082.54,"unite":"Elo","bas":1071.54,"haut":1093.54,"votes":4600,"chiffres":{"prix_par_million_caracteres":15,"date_sortie":"2025-06-03"},"voisins":["google-gemini-2-5-flash-lite-tts","google-studio"],"source":{"nom":"Artificial Analysis","mention":"Source : Artificial Analysis","url":"https://artificialanalysis.ai/text-to-speech/model-families/fishaudio","releve":"2026-10-19"}},{"id":"synthese/tts-arena","nom":"Synthèse vocale en anglais, TTS Arena","rang":24,"sur":38,"valeur":1504,"unite":"Elo","bas":1473,"haut":1535,"votes":538,"voisins":["cartesia-sonic-2","minimax-speech-2-6-hd"],"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":{"artificial_analysis":"OpenAudio S1","tts_arena":"OpenAudio S1"}}