Vendor page
omniASR-LLM-7B model card (Omnilingual ASR)
Meta FAIR, Hugging Face, 2025
Read the originalRead on October 17, 2026
- What it is
- Meta's model card for the 7-billion-parameter Omnilingual ASR model, which the card describes as open-source speech recognition for more than 1,600 languages. It gives the license for the model, the code and the training corpus.
- How much weight it carries
- Written by the model's publisher, so it is authoritative for license and language coverage. Meta reports that accuracy varies widely across its languages, so the language count is not a promise of quality in each one.
Facts taken from this source
- Meta's Omnilingual ASR is an open-source speech recognition model family for more than 1,600 languages.
“Omnilingual ASR: Open-Source Multilingual Speech Recognition for 1600+ Languages” (Page title)
- Meta's Omnilingual ASR model and its code are licensed under Apache 2.0, while the training corpus is licensed separately under CC-BY-4.0.
“License: Apache-2.0 (for the model and code), CC-BY-4.0 for the facebook/omnilingual-asr-corpus dataset.” (License line, model description)