whatmodel.sh / stt / Canary-1B-Flash

Canary-1B-Flash

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

Canary-1B-Flash is an open-weights speech-to-text model from NVIDIA, released March 2025. Its published size is 0.81B parameters. Licence: cc-by-4.0.

Catalogue identity notes

Exact checkpoint nvidia/canary-1b-flash (NeMo). Params: HF safetensors total (card text says 883M). released: HF repo creation month.

59.9±2.1#33 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-Flash, out of 176 with a score, then more from NVIDIA.

#modelscorewhereRAMreleased
38Grok Voice Transcribe 1.0 xAI60±5comparison only2026-04
39Qwen3-ASR-1.7B Alibaba (Qwen)60±1api $0.0005/audio min +local~3.3GB2026-01
40Canary-1B-Flash NVIDIA60±2local~4.4GB2025-03
41Granite Speech 3.3 2B IBM60±3local~7.7GB2025-06
42Speechmatics Enhanced Speechmatics60±2api $0.0067/audio min2023-03
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 tests5 not in a task · 98% of its score
testresultvs the others in its rundifficultycountswhy
AfriSpeech-MultiBench · 7-region African-accent mean WER23.8% ✎▂····▂···▂▆█ beats 5/11avg 46%▇▇▇▇▇▇▇▇▇▇▇▇ 29%typical
AfriSpeech-MultiBench · Parliament (legal domain) WER19.1% ✎▂▂····▂·▅·▃█ beats 9/12avg 52%▇▇▇▇▇▇▇▇▇▇ 25%sharp
LibriSpeech test-clean WER1.6% ✎▁···▁▁▂·▄▄██ beats 42/56avg 2.3%▇▇▇▇▇▇▇▇ 20%blunt
Open ASR English short-form WER (March 2026)6.3% ✎▁▁▁▃··▁▁▂██▆ beats 73/86avg 8.2%▇▇▇▇▇▇ 16%blunt
FLEURS English test WER4.8% ✎▁··▁▁▁▁▁▁▃█▁ beats 30/56avg 5.8%▇▇▇ 8.0%blunt

$ where to run it

Other ways to run it

  1. Run it as published · needs ~4.39 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-flash').transcribe(['audio.wav']))"

Official weights and documentation: huggingface.co/nvidia ↗

$ how the score was worked out · AfriSpeech-MultiBench (arXiv:… 55%, transcribe.cpp model catalog… 28%, Open ASR Leaderboard 16%

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
AfriSpeech-MultiBench (arXiv:2511.14255v1) · African-accented English ASR benchmark suite2▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 55%CC BY 4.0
transcribe.cpp model catalog · LibriSpeech test-clean WER of validated full-precision ports2▇▇▇▇▇▇▇▇ 28%MIT
Open ASR Leaderboard1▇▇▇▇▇ 16%Apache-2.0
prior · pull toward the middle▇ 2%
weighted average= 59.9

$ 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