{"id":"meta-muse-voice-transcribe","nom":"muse-voice-transcribe","editeur":"Meta","pays":"États-Unis","poids_ouverts":false,"poids_selon":"Open ASR Leaderboard","licence":null,"licence_selon":null,"liens":{"page":"https://developer.meta.com/ai/models/muse-voice-transcribe/","essayer":"https://research.meta.ai/blog/introducing-muse-voice-transcribe"},"fiche_auto":true,"page":"https://quelleia.com/voix/modeles/meta-muse-voice-transcribe/","releve":"2026-10-19","en_bref":"26e sur 39 en transcription en français ; 28e sur 53 en transcription en anglais (Open ASR).","presentation":"Modèle de transcription en temps réel de Meta Superintelligence Labs, proposé par la Meta Model API. Il réunit reconnaissance en flux, attribution de la parole à plus de 20 locuteurs et détection de fin de parole, et gère le passage d'une langue à l'autre.","sortie":"2026-09-01","langues":{"annoncees":{"annoncees":"25 langues validées, dont le français (entraîné sur plus de 70)","nombre":25,"francais":true,"liste":["ar","bn","nl","en","fr","de","he","hi","id","it","ja","kn","ko","ms","zh","mr","pl","pt","es","tl","ta","te","th","tr","vi"]},"prises_en_charge_selon_open_tts":null,"evaluees":["fr","en"]},"preuves":[{"appuie":["sortie"],"lu":"https://research.meta.ai/blog/introducing-muse-voice-transcribe (« Introducing Muse Voice Transcribe September 1, 2026 »), 17/10/2026"},{"appuie":["presentation"],"lu":"https://research.meta.ai/blog/introducing-muse-voice-transcribe (« We’re excited to introduce Muse Voice Transcribe, the first real-time audio perception model developed by Meta Superintelligence Labs. »), 17/10/2026"},{"appuie":["presentation"],"lu":"https://research.meta.ai/blog/introducing-muse-voice-transcribe (« It delivers real-time streaming ASR, diarization with 20+ speakers, and endpointing. »), 17/10/2026"},{"appuie":["langues"],"lu":"https://research.meta.ai/blog/introducing-muse-voice-transcribe (« Muse Voice Transcribe is trained with 70+ languages, of which 25 are extensively verified. »), 17/10/2026"},{"appuie":["langues"],"lu":"https://dev.meta.ai/docs/speech-to-text/ (« Muse Voice Transcribe supports 25 languages with code-switching: Arabic, Bengali, Dutch, English, French, German, Hebrew, Hindi, »), 17/10/2026"},{"appuie":["presentation"],"lu":"https://dev.meta.ai/docs/models/ (« Muse Voice Transcribe is Meta's speech-to-text model on Meta Model API. »), 17/10/2026"}],"classements":[{"id":"transcription/francais","nom":"Transcription en français","rang":26,"sur":39,"valeur":6.81,"unite":"%","chiffres":{"jeux":{"Common Voice":9.41,"FLEURS":5.94,"MLS":5.09}},"voisins":["whisper-large-v3-turbo","meta-omniasr-llm-3b-v2"],"source":{"nom":"Open ASR Leaderboard","mention":"Source : Open ASR Leaderboard, Hugging Face","url":"https://huggingface.co/spaces/hf-audio/open_asr_leaderboard","releve":"2026-10-19"}},{"id":"transcription/open-asr-anglais","nom":"Transcription en anglais, Open ASR","rang":28,"sur":53,"valeur":6.3,"unite":"%","chiffres":{"jeux":{"AMI-Cleaned":9.2,"Earnings22-Cleaned-AA-chunked":6.01,"Gigaspeech-Cleaned":7.75,"LS Clean":1.84,"LS Other":4.51,"Private (conversational)":13.04,"Private (scripted)":4.52,"SPGISpeech":3.21,"URGENT24":13.03,"Voice Arena Monsoon":4.16,"Voxpopuli-AA-Cleaned":2.02}},"voisins":["alibaba-qwen3-asr-0-6b-hf","ibm-granite-speech-5-0-470m-turboctc"],"source":{"nom":"Open ASR Leaderboard","mention":"Source : Open ASR Leaderboard, Hugging Face","url":"https://huggingface.co/spaces/hf-audio/open_asr_leaderboard","releve":"2026-10-19"}}],"noms_dans_les_sources":{"hf_audio":"meta/muse-voice-transcribe"}}