whatmodel.sh / stt / Parakeet CTC 0.6B

Parakeet CTC 0.6B

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

Parakeet CTC 0.6B is an open-weights speech-to-text model from NVIDIA, released January 2024. Its published size is 0.61B parameters. Licence: cc-by-4.0.

Catalogue identity notes

Exact checkpoint nvidia/parakeet-ctc-0.6b (NeMo). Params: HF safetensors total. released aligned with catalogue sibling parakeet-rnnt-1-1b (repo created 2023-12-28).

54.0±2.7#79 of 162 speech-to-text modelsfrom 4 benchmark results across 2 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 Parakeet CTC 0.6B, out of 176 with a score, then more from NVIDIA.

#modelscorewhereRAMreleased
88Ink-Whisper Cartesia55±3api price unknown2025-06
89Breeze ASR 25 MediaTek Research54±4local~1.7GB2025-06
90Parakeet CTC 0.6B NVIDIA54±3local~3.4GB2024-01
91Canary-180M-Flash NVIDIA54±3local~1.4GB2025-03
92Omnilingual ASR CTC 3B Meta54±2local~7.9GB2025-11
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

long-form transcription▁▁▄▁·▁▂█▄▁▁▁ #16 of 27 * · 26% of its score · 1 test

Beats Parakeet TDT 1.1B, Parakeet RNNT 1.1B, Distil-Whisper distil-large-v3 here, though they rank higher overall. Trails Whisper large (v1), Whisper large-v2, Rev AI Machine (Reverb ASR) here, though they rank lower overall.

testresultvs the others in its rundifficultycountswhy
Open ASR English long-form WER9.8% ✎▁▁▅▁·▃▅█▆▁▂▂ beats 12/30avg 12%▇▇▇▇▇▇▇▇▇▇ 26%typical
other tests3 not in a task · 72% of its score
testresultvs the others in its rundifficultycountswhy
LibriSpeech test-clean WER1.9% ✎▁···▁▁▂·▄▄██ beats 32/56avg 2.3%▇▇▇▇▇▇▇▇▇▇▇▇▇ 33%blunt
Open ASR English short-form WER (March 2026)7.7% ✎▁▁▁▃··▁▁▂██▆ beats 44/86avg 8.2%▇▇▇▇▇▇▇▇▇▇ 26%blunt
FLEURS English test WER5.5% ✎▁··▁▁▁▁▁▁▃█▁ beats 20/56avg 5.8%▇▇▇▇▇ 13%blunt

$ where to run it

Other ways to run it

  1. Run it as published · needs ~3.43 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/parakeet-ctc-0.6b').transcribe(['audio.wav']))"

Official weights and documentation: huggingface.co/nvidia ↗

$ how the score was worked out · Open ASR Leaderboard 51%, transcribe.cpp model catalog… 46%

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 Leaderboard2▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 51%Apache-2.0
transcribe.cpp model catalog · LibriSpeech test-clean WER of validated full-precision ports2▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 46%MIT
prior · pull toward the middle▇ 3%
weighted average= 54.0

$ 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