whatmodel.sh / stt / Canary-1B

Canary-1B

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

Canary-1B is an open-weights speech-to-text model from NVIDIA, released February 2024. Its published size is 1B parameters. Licence: cc-by-nc-4.0.

Catalogue identity notes

Exact checkpoint nvidia/canary-1b (NeMo). Original 2024 Canary-1B; distinct from catalogue canary-1b-v2 and canary-1b-flash. Weights CC-BY-NC-4.0. Params: card ('1 billion'). released: HF repo creation month.

58.9±3.1#40 of 162 speech-to-text modelsfrom 5 benchmark results across 3 sources

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 Canary-1B, out of 176 with a score, then more from NVIDIA.

#modelscorewhereRAMreleased
47Avalon v1 (en) Aqua Voice59±5comparison only2025-08
48GPT-Realtime-Whisper OpenAI59±3api $0.017/audio min2026-05
49Canary-1B NVIDIA59±3local~5.3GB2024-02
50Parakeet Ultra Moondream59±3local~0.7GB2026-09
51Parakeet TDT 1.1B NVIDIA59±2local~5.8GB2024-01
more from NVIDIA
28Canary-Qwen-2.5B NVIDIA61±2local~6.6GB2025-07
35Parakeet TDT 0.6B v2 NVIDIA60±1local~1.9GB2025-05
37Parakeet Unified EN 0.6B NVIDIA60±4local~0.7GB2026-04

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

$ what it's good at

other tests4 not in a task · 96% of its score
testresultvs the others in its rundifficultycountswhy
LibriSpeech test-clean WER +11.6% ✎▁···▁▁▂·▄▄██ beats 46/56avg 2.3%▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 44%blunt · several runs
Open ASR English short-form WER (March 2026)6.5% ✎▁▁▁▃··▁▁▂██▆ beats 69/86avg 8.2%▇▇▇▇▇▇▇▇▇▇▇▇▇ 33%blunt
FLEURS English test WER4.4% ✎▁··▁▁▁▁▁▁▃█▁ beats 40/56avg 5.8%▇▇▇▇▇▇▇ 17%blunt
TED-LIUM test WER (short-form)10.2% ✎▂····▂▂▂█▃▃▃ beats 13/13avg 9.2%▇ 2.4%blunt

$ where to run it

Other ways to run it

  1. Run it as published · needs ~5.3 GB, NVIDIA only · from huggingface.co/nvidia (official)
    uvx --with 'nemo_toolkit[asr]' python -c "import nemo.collections.asr as a; print(a.models.ASRModel.from_pretrained('nvidia/canary-1b').transcribe(['audio.wav']))"

Official weights and documentation: huggingface.co/nvidia ↗

$ how the score was worked out · transcribe.cpp model catalog… 59%, Open ASR Leaderboard 33%, Speech Robust Bench (arXiv:24… 4%

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
transcribe.cpp model catalog · LibriSpeech test-clean WER of validated full-precision ports2▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 59%MIT
Open ASR Leaderboard1▇▇▇▇▇▇▇▇▇▇ 33%Apache-2.0
Speech Robust Bench (arXiv:2403.07937v3) · clean-speech WER of 25+ open ASR models2▇ 4%CC BY 4.0
prior · pull toward the middle▇ 4%
weighted average= 58.9

$ 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