{"id":"cohere-transcribe","nom":"Cohere Transcribe","editeur":"Cohere","pays":"Canada","poids_ouverts":true,"poids_selon":null,"licence":"Apache 2.0","licence_selon":null,"liens":{"page":"https://cohere.com/blog/transcribe","essayer":"https://huggingface.co/spaces/CohereLabs/cohere-transcribe-03-2026"},"fiche_auto":false,"page":"https://quelleia.com/voix/modeles/cohere-transcribe/","releve":"2026-10-19","en_bref":"8e sur 39 en transcription en français ; 37e sur 51 en transcription en anglais (Artificial Analysis) ; 11e sur 53 en transcription en anglais (Open ASR).","presentation":"Premier modèle de transcription de Cohere, publié avec ses poids et servi aussi par son API, qui ne fait que convertir la parole en texte. Il repose sur un grand encodeur Conformer suivi d'un décodeur Transformer léger, entraîné de zéro.","sortie":"2026-03-26","taille":"2 milliards de paramètres","langues":{"annoncees":{"annoncees":"14 langues, dont le français","nombre":14,"francais":true,"liste":["en","de","fr","it","es","pt","el","nl","pl","vi","zh","ar","ja","ko"]},"prises_en_charge_selon_open_tts":null,"evaluees":["fr","en"]},"preuves":[{"appuie":["presentation","sortie"],"lu":"https://docs.cohere.com/changelog/cohere-transcribe-03-2026 (« March 26, 2026 Announcing the Cohere Transcribe model We’re pleased to announce the release of Cohere Transcribe , our first transcription model. »), 17/10/2026"},{"appuie":["presentation"],"lu":"https://docs.cohere.com/changelog/cohere-transcribe-03-2026 (« The model is available immediately through Cohere’s Audio Transcriptions API endpoint . »), 17/10/2026"},{"appuie":["sortie"],"lu":"https://cohere.com/blog/transcribe (« Mar 26, 2026 3 minute read Introducing Cohere Transcribe: a new state-of-the-art in open-source speech recogniti »), 17/10/2026"},{"appuie":["langues","presentation","taille"],"lu":"https://huggingface.co/CohereLabs/cohere-transcribe-03-2026 (« Cohere Transcribe is an open source release of a 2B parameter dedicated audio-in, text-out automatic speech recognition (ASR) model. The model supports 14 languages. »), 17/10/2026"},{"appuie":["langues"],"lu":"https://huggingface.co/CohereLabs/cohere-transcribe-03-2026 (« Trained on 14 languages: European: English, French, German, Italian, Spanish, Portuguese, Greek, Dutch, Polish APAC: Chinese (Mandarin), Japanese, Korean, Vietnamese MENA: Arabic License Apache 2.0 »), 17/10/2026"},{"appuie":["presentation"],"lu":"https://huggingface.co/CohereLabs/cohere-transcribe-03-2026 (« Training objective supervised cross-entropy on output tokens; trained from scratch »), 17/10/2026"},{"appuie":["presentation"],"lu":"https://huggingface.co/CohereLabs/cohere-transcribe-03-2026 (« Model a large Conformer encoder extracts acoustic representations, followed by a lightweight Transformer decoder for token generation »), 17/10/2026"}],"classements":[{"id":"transcription/francais","nom":"Transcription en français","rang":8,"sur":39,"valeur":4.02,"unite":"%","chiffres":{"rtfx":635.11,"jeux":{"Common Voice":5.28,"FLEURS":4.33,"MLS":2.44}},"voisins":["assemblyai-universal-3-5-pro","resonant-1"],"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":37,"sur":51,"valeur":4.57,"unite":"%","chiffres":{"prix_pour_1000_minutes":0,"jeux":{"AA-AgentTalk":3.09,"Earnings22":6.73,"VoxPopuli":5.36},"date_sortie":"2026-03-01","hebergeur":"Cohere"},"voisins":["openai-gpt-4o-mini-transcribe","whisper-large-v3-turbo"],"source":{"nom":"Artificial Analysis","mention":"Source : Artificial Analysis","url":"https://artificialanalysis.ai/speech-to-text/models/cohere-transcribe","releve":"2026-10-19"}},{"id":"transcription/open-asr-anglais","nom":"Transcription en anglais, Open ASR","rang":11,"sur":53,"valeur":5.77,"unite":"%","chiffres":{"rtfx":902.46,"jeux":{"AMI-Cleaned":7.02,"Earnings22-Cleaned-AA-chunked":7.88,"Gigaspeech-Cleaned":7.88,"LS Clean":0.97,"LS Other":2.05,"Private (conversational)":13.46,"Private (scripted)":3.2,"SPGISpeech":2.74,"URGENT24":9.5,"Voice Arena Monsoon":3.44,"Voxpopuli-AA-Cleaned":5.38}},"voisins":["smallest-ai-pulse","sprag-symphony"],"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":"CohereLabs/cohere-transcribe-03-2026","artificial_analysis":"Cohere Transcribe"}}