whatmodel.sh / stt / Parakeet Unified EN 0.6B
Parakeet Unified EN 0.6B
↗ share
+ contribute
Parakeet Unified EN 0.6B is an open-weights speech-to-text model from NVIDIA, released April 2026. Its published size is 0.618B parameters. Licence: nvidia-open-model-license.
Catalogue identity notes
nvidia/parakeet-unified-en-0.6b: English FastConformer RNN-T model for both offline and streaming inference (card: 'combining both offline and streaming inference (with a minimum latency of 160ms) in one model'; 'Parameters: 600M'). Not the cache-aware nemotron-speech-streaming-en-0.6b or parakeet-tdt-0.6b-v2. transcribe.cpp port of upstream commit d4ac992 (2026-04-22); the only later commit fe53cd8 (2026-06-09) edits README.md. Repo created 2026-04-07.
Model information sources
Release date ↗ · Parameter count ↗
transcribe.cpp catalog 'params' 618,309,121 for its converted port of nvidia/parakeet-unified-en-0.6b@d4ac992 (no safetensors index on the Hugging Face repo).
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 Unified EN 0.6B, out of 176 with a score, then more from NVIDIA.
| # | model | score | where | ||
|---|---|---|---|---|---|
| 35 | Parakeet TDT 0.6B v2 | 60±1 | local | ||
| 36 |  Tencent Hy-ASR-3.0-preview | 60±4 | comparison only | ||
| 37 | Parakeet Unified EN 0.6B | 60±4 | local | ||
| 38 |  Grok Voice Transcribe 1.0 | 60±5 | comparison only | ||
| 39 |  Qwen3-ASR-1.7B | 60±1 | api $0.0005/audio min +local | ||
| more from NVIDIA | |||||
| 28 | Canary-Qwen-2.5B | 61±2 | local | ||
| 40 | Canary-1B-Flash | 60±2 | local | ||
| 49 | Canary-1B | 59±3 | local | ||
see all speech-to-text models · find the ones that fit your machine
$ what it's good at
other tests2 not in a task · 94% of its score
| test | result | vs the others in its run | counts |
|---|---|---|---|
| LibriSpeech test-clean WER | 1.6% ✎ | ▁···▁▁▂·▄▄██ beats 44/56 | 67% |
| FLEURS English test WER | 4.0% ✎ | ▁··▁▁▁▁▁▁▃█▃ beats 52/56 | 27% |
$ where to run it
Other ways to run it
- Run it as published · needs ~0.73 GB · from github.com/handy-computer
git clone https://github.com/handy-computer/transcribe.cpp.git && cd transcribe.cpp && cmake -B build && cmake --build build && hf download handy-computer/parakeet-unified-en-0.6b-gguf parakeet-unified-en-0.6b-Q8_0.gguf --local-dir models/parakeet-unified-en-0.6b && build/bin/transcribe-cli -m models/parakeet-unified-en-0.6b/parakeet-unified-en-0.6b-Q8_0.gguf audio.wav
Official weights and documentation: huggingface.co/nvidia ↗
$ how the score was worked out · transcribe.cpp model catalog… 94%
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
| source | results | share of the score |
|---|---|---|
 transcribe.cpp model catalog · LibriSpeech test-clean WER of validated full-precision ports | 2 | 94% |
| prior · pull toward the middle | 6% | |
| weighted average | = 60.4 |
$ improve this page
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