{"id":"meta-omniasr-llm-300m-v2","nom":"omniASR-LLM-300M-v2","editeur":"Meta","pays":"États-Unis","poids_ouverts":true,"poids_selon":null,"licence":"Apache 2.0","licence_selon":null,"liens":{"page":"https://github.com/facebookresearch/omnilingual-asr"},"fiche_auto":true,"page":"https://quelleia.com/voix/modeles/meta-omniasr-llm-300m-v2/","releve":"2026-10-19","en_bref":"34e sur 39 en transcription en français.","presentation":"Modèle ouvert de reconnaissance vocale de la famille Omnilingual ASR de Meta, version v2 de décembre 2025 qui associe l'encodeur audio à un décodeur de type LLM. On peut lui indiquer la langue attendue, et il couvre plus de 1600 langues.","sortie":"2025-12","taille":"1,63 milliard de paramètres","langues":{"annoncees":{"annoncees":"plus de 1600 langues, dont le français","nombre":1600,"plus":true,"francais":true},"prises_en_charge_selon_open_tts":null,"evaluees":["fr"]},"preuves":[{"appuie":["sortie"],"lu":"https://raw.githubusercontent.com/facebookresearch/omnilingual-asr/main/README.md (« ## December 2025 Update »), 17/10/2026"},{"appuie":["presentation","sortie"],"lu":"https://raw.githubusercontent.com/facebookresearch/omnilingual-asr/main/README.md (« We release two suites of models: - Checkpoints of improved accuracy (CER) for the CTC and LLM-ASR models compared to our existing LLM-ASR model (`omniASR_{CTC,LLM}_{300M,1B,3B,7B}_v2`). »), 17/10/2026"},{"appuie":["langues","presentation"],"lu":"https://raw.githubusercontent.com/facebookresearch/omnilingual-asr/main/README.md (« Omnilingual ASR is an open-source speech recognition system supporting over 1,600 languages »), 17/10/2026"},{"appuie":["langues"],"lu":"https://raw.githubusercontent.com/facebookresearch/omnilingual-asr/main/per_language_results_table_7B_llm_asr.csv (« fra_Latn,4615.1,2.2 »), 17/10/2026"},{"appuie":["presentation","taille"],"lu":"https://raw.githubusercontent.com/facebookresearch/omnilingual-asr/main/README.md (« | [`omniASR_LLM_300M_v2`](https://dl.fbaipublicfiles.com/mms/omniASR-LLM-300M-v2.pt) | ASR with optional language conditioning | 1_627_603_584 | »), 17/10/2026"}],"classements":[{"id":"transcription/francais","nom":"Transcription en français","rang":34,"sur":39,"valeur":10.02,"unite":"%","chiffres":{"rtfx":54.82,"jeux":{"Common Voice":12.07,"FLEURS":11.21,"MLS":6.78}},"voisins":["nvidia-nemotron-3-5-asr-streaming-0-6b","meta-omniasr-ctc-1b-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"}}],"noms_dans_les_sources":{"hf_audio":"facebook/omniASR-LLM-300M-v2"}}