whatmodel.sh / stt / Omnilingual ASR CTC 3B v2

Omnilingual ASR CTC 3B v2

+ contribute
Update model informationAdd a place to run itAdd a test resultPropose a benchmarkCorrect a benchmark scoreEdit data on GitHub

Omnilingual ASR CTC 3B v2 is an open-weights speech-to-text model from Meta, released December 2025. Its published size is 3.08B parameters. Licence: apache-2.0.

Catalogue identity notes

Omnilingual ASR model card 'omniASR_CTC_3B_v2' (checkpoint https://dl.fbaipublicfiles.com/mms/omniASR-CTC-3B-v2.pt, listed in src/omnilingual_asr/cards/models/rc_models_v2.yaml @ 81f51e2; 'December 2025 Update'). Not on Hugging Face; the CSV entry 'facebook/omniASR-CTC-3B-v2' is the leaderboard's label. Distinct from the v1 checkpoint of the same size. Params: README table. Needs libsndfile (README).

55.0±2.9#72 of 162 speech-to-text modelsfrom 1 benchmark result across 1 source* thinly tested, treat the rank as provisional

Score within the Open ASR English short-form WER (March 2026) comparison set only.

$ how it compares

The speech-to-text models ranked either side of Omnilingual ASR CTC 3B v2, out of 176 with a score, then more from Meta.

#modelscorewhereRAMreleased
81Scribe v2 Realtime ElevenLabs55±3api $0.0065/audio min2025-11
82LiteWhisper large-v3-acc Efficient-Speech (LiteASR)55±3local~3.9GB2025-02
83Omnilingual ASR CTC 3B v2 Meta55±3local~7.9GB2025-12
84Omnilingual ASR CTC 7B Meta55±2local~16GB2025-11
85Parakeet CTC 1.1B NVIDIA55±3local~5.6GB2024-01
more from Meta
1Muse Voice Transcribe 1.0 Meta71±3api $0.0030/audio min2026-09
58Omnilingual ASR LLM 7B Meta58±2local~19GB2025-11
59Omnilingual ASR LLM 7B v2 Meta58±3local~19GB2025-12

see all speech-to-text models · find the ones that fit your machine

$ what it's good at

multilingual transcription▂··▃▄█▆▅▄▂·▂ #22 of 36 * · 97% of its score · 1 test

Beats Voxtral Mini Transcribe V2, Canary-1B-v2, Omnilingual ASR LLM 3B here, though they rank higher overall. Trails Omnilingual ASR CTC 1B v2, Omnilingual ASR CTC 7B v2, SeamlessM4T v2 Large here, though they rank lower overall.

testresultvs the others in its rundifficultycountswhy
Open ASR multilingual WER6.2% ✎▂··▂▃▂█▃█▂▁▁ beats 12/29avg 8.9%▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 97%typical

$ where to run it

Other ways to run it

  1. Run it as published · needs ~7.89 GB · from github.com/facebookresearch (official)
    uv run --python 3.12 --with omnilingual-asr python -c "from omnilingual_asr.models.inference.pipeline import ASRInferencePipeline as P; print(P(model_card='omniASR_CTC_3B_v2').transcribe(['audio.wav'], batch_size=1))"

Official weights and documentation: github.com/facebookresearch ↗

$ how the score was worked out · Open ASR Leaderboard 97%

Each result implies a score on its own, after adjusting for how hard its test is and who ran it. The score is their weighted average, with a small pull toward the middle for thinly tested models. Each test's weight is its counts column above; here it's added up by who ran the tests. How it works

sourceresultsshare of the scorelicence
Open ASR Leaderboard1▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 97%Apache-2.0
prior · pull toward the middle▇ 3%
weighted average= 55.0

$ improve this page

Update model informationAdd a place to run itAdd a test resultPropose a benchmarkCorrect a benchmark scoreEdit data on GitHub

Use a short GitHub form or edit the data through a pull request. Every number needs a source; here's which ones we can use.

data updated · sources and licences · json