{"id":"microsoft-azure-speech-07-2026","nom":"azure-speech-07-2026","editeur":"Microsoft","pays":"États-Unis","poids_ouverts":false,"poids_selon":"Open ASR Leaderboard","licence":null,"licence_selon":null,"liens":{"page":"https://learn.microsoft.com/en-us/azure/ai-services/speech-service/llm-speech"},"fiche_auto":true,"page":"https://quelleia.com/voix/modeles/microsoft-azure-speech-07-2026/","releve":"2026-10-19","en_bref":"1er sur 39 en transcription en français ; 3e sur 53 en transcription en anglais (Open ASR).","presentation":"Service de transcription d'Azure AI Speech (API LLM Speech de Microsoft Foundry), où un grand modèle de langage renforce un modèle de reconnaissance vocale. Il détecte la langue automatiquement, sait aussi traduire l'audio et accepte des consignes (prompt) pour orienter le style de sortie.","langues":{"annoncees":{"annoncees":"25 langues, dont le français","nombre":25,"francais":true,"liste":["ar","zh","cs","da","nl","en","fi","fr","de","el","he","hi","hu","id","it","ja","ko","nb","pl","pt","ru","es","sv","th","tr"]},"prises_en_charge_selon_open_tts":null,"evaluees":["fr","en"]},"preuves":[{"appuie":["presentation"],"lu":"https://learn.microsoft.com/en-us/azure/ai-services/speech-service/llm-speech (« A large language model (LLM) enhances a speech model, delivering improved quality, deep contextual understanding, multilingual support, and prompt-tuning capabilities. »), 17/10/2026"},{"appuie":["presentation"],"lu":"https://learn.microsoft.com/en-us/azure/ai-services/speech-service/llm-speech (« The service operates in multi-lingual mode by default, so you don't need to specify the input language locale. »), 17/10/2026"},{"appuie":["langues"],"lu":"https://learn.microsoft.com/en-us/azure/ai-services/speech-service/llm-speech (« The following input languages are supported for both transcribe and translate tasks: Arabic , Chinese , Czech , Danish , Dutch , English , Finnish , French , German »), 17/10/2026"},{"appuie":["presentation"],"lu":"https://devblogs.microsoft.com/foundry/announcing-azure-ai-speech-llm-2607/ (« This release delivers a decisive quality leap across Tier 1 and Tier 2 languages, positioning our model at #2 standing on HuggingFace’s OpenASR leaderboard »), 17/10/2026"}],"classements":[{"id":"transcription/francais","nom":"Transcription en français","rang":1,"sur":39,"valeur":3.14,"unite":"%","chiffres":{"jeux":{"Common Voice":4.32,"FLEURS":2.48,"MLS":2.62}},"voisins":["microsoft-azure-speech-06-2026"],"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":3,"sur":53,"valeur":5.02,"unite":"%","chiffres":{"jeux":{"AMI-Cleaned":7.05,"Earnings22-Cleaned-AA-chunked":5.2,"Gigaspeech-Cleaned":7.48,"LS Clean":1.23,"LS Other":2.71,"Private (conversational)":12.71,"Private (scripted)":2.12,"SPGISpeech":2.17,"URGENT24":9.9,"Voice Arena Monsoon":3,"Voxpopuli-AA-Cleaned":1.65}},"voisins":["zoom-scribe-v2-pro","modulate-multilingual"],"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":"microsoft/azure-speech-07-2026"}}