{"id":"assemblyai-universal-3-5-pro","nom":"Universal-3.5 Pro","editeur":"AssemblyAI","pays":"États-Unis","poids_ouverts":false,"poids_selon":"Open ASR Leaderboard","licence":null,"licence_selon":null,"liens":{"page":"https://www.assemblyai.com/docs/pre-recorded-audio/universal-3-5-pro","essayer":"https://www.assemblyai.com/playground"},"fiche_auto":false,"page":"https://quelleia.com/voix/modeles/assemblyai-universal-3-5-pro/","releve":"2026-10-19","en_bref":"6e sur 39 en transcription en français (ex æquo à valeur égale) ; 13e sur 51 en transcription en anglais (Artificial Analysis) ; 8e sur 53 en transcription en anglais (Open ASR).","presentation":"Universal-3.5 Pro est le modèle de transcription principal d'AssemblyAI, en différé et en temps réel, qui suit les changements de langue sur 18 langues et distingue les orateurs. Il accepte un contexte rédigé (prompting) et jusqu'à 1 000 termes clés, et l'éditeur le présente comme son modèle le plus rapide.","sortie":"2026-07-07","langues":{"annoncees":{"annoncees":"18 langues, dont le français","nombre":18,"francais":true,"liste":["en","es","de","fr","pt","it","tr","nl","sv","no","da","fi","hi","vi","ar","he","ja","zh"]},"prises_en_charge_selon_open_tts":null,"evaluees":["fr","en"]},"preuves":[{"appuie":["sortie"],"lu":"https://www.assemblyai.com/blog/universal-3-5-pro-async (« Published on 7 July 2026 »), 17/10/2026"},{"appuie":["presentation"],"lu":"https://www.assemblyai.com/blog/universal-3-5-pro-async (« Today we're releasing Universal-3.5 Pro, our new flagship async speech-to-text model. »), 17/10/2026"},{"appuie":["langues","presentation"],"lu":"https://www.assemblyai.com/blog/universal-3-5-pro-async (« Native code-switching across 18 languages captures every word in the language it was spoken. »), 17/10/2026"},{"appuie":["presentation"],"lu":"https://www.assemblyai.com/docs/getting-started/universal-3-5-pro.md (« Universal-3.5 Pro is available for both **pre-recorded (async)** and **streaming** use cases. »), 17/10/2026"},{"appuie":["presentation"],"lu":"https://www.assemblyai.com/docs/getting-started/universal-3-5-pro.md (« It's also our fastest model, so you get the best accuracy without sacrificing speed. »), 17/10/2026"},{"appuie":["presentation"],"lu":"https://www.assemblyai.com/docs/getting-started/models.md (« Keyterms prompting up to 1,000 words »), 17/10/2026"},{"appuie":["langues"],"lu":"https://www.assemblyai.com/docs/getting-started/models.md (« Supports 18 languages »), 17/10/2026"},{"appuie":["langues"],"lu":"https://www.assemblyai.com/docs/pre-recorded-audio/supported-languages.md (« Universal-3.5 Pro supports the following 18 languages. »), 17/10/2026"},{"appuie":["langues"],"lu":"https://www.assemblyai.com/docs/pre-recorded-audio/supported-languages.md (« Whether your audio features Quebecois French, Mexican Spanish, or Brazilian Portuguese, the model accurately captures speech as it's naturally spoken »), 17/10/2026"}],"classements":[{"id":"transcription/francais","nom":"Transcription en français","rang":6,"sur":39,"valeur":3.92,"unite":"%","chiffres":{"jeux":{"Common Voice":6.47,"FLEURS":2.82,"MLS":2.47}},"voisins":["sprag-symphony","cohere-transcribe"],"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/artificial-analysis","nom":"Transcription en anglais, indice AA-WER","rang":13,"sur":51,"valeur":2.99,"unite":"%","chiffres":{"prix_pour_1000_minutes":3.5,"jeux":{"AA-AgentTalk":1.94,"Earnings22":5.95,"VoxPopuli":2.13},"date_sortie":"2026-06-23","hebergeur":"AssemblyAI"},"voisins":["google-gemini-3-flash-high","universal-3-pro"],"source":{"nom":"Artificial Analysis","mention":"Source : Artificial Analysis","url":"https://artificialanalysis.ai/speech-to-text/models/assemblyai","releve":"2026-10-19"}},{"id":"transcription/open-asr-anglais","nom":"Transcription en anglais, Open ASR","rang":8,"sur":53,"valeur":5.39,"unite":"%","chiffres":{"jeux":{"AMI-Cleaned":10.63,"Earnings22-Cleaned-AA-chunked":6.05,"Gigaspeech-Cleaned":7.62,"LS Clean":1.1,"LS Other":2.11,"Private (conversational)":12.8,"Private (scripted)":2.66,"SPGISpeech":1.81,"URGENT24":9.08,"Voice Arena Monsoon":3.49,"Voxpopuli-AA-Cleaned":1.92}},"voisins":["scribe-v2","alibaba-qwen3-asr-1-7b-hf"],"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":"assemblyai/universal-3-5-pro","artificial_analysis":"Universal-3.5 Pro"}}