whatmodel.sh / stt / Parakeet TDT 1.1B
Parakeet TDT 1.1B
↗ share
+ contribute
Parakeet TDT 1.1B is an open-weights speech-to-text model from NVIDIA, released January 2024. Its published size is 1.1B parameters. Licence: cc-by-4.0.
Catalogue identity notes
Exact checkpoint nvidia/parakeet-tdt-1.1b (NeMo). Params: card (~1.1B). released: HF repo creation month.
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 TDT 1.1B, out of 176 with a score, then more from NVIDIA.
| # | model | score | where | ||
|---|---|---|---|---|---|
| 49 | Canary-1B | 59±3 | local | ||
| 50 |  Parakeet Ultra | 59±3 | local | ||
| 51 | Parakeet TDT 1.1B | 59±2 | local | ||
| 52 |  Voxtral Mini 3B (2507) | 58±1 | api $0.0010/audio min +local | ||
| 53 |  Universal-Streaming (English) | 58±3 | api $0.0025/audio min | ||
| more from NVIDIA | |||||
| 28 | Canary-Qwen-2.5B | 61±2 | local | ||
| 35 | Parakeet TDT 0.6B v2 | 60±1 | local | ||
| 37 | Parakeet Unified EN 0.6B | 60±4 | local | ||
see all speech-to-text models · find the ones that fit your machine
$ what it's good at
long-form transcription▁▁▄▁·▁▃█▄▁▁▁ #18 of 27 * · 13% of its score · 1 test
Trails Whisper large (v1), Whisper large-v2, Rev AI Machine (Reverb ASR) here, though they rank lower overall.
| test | result | vs the others in its run | counts |
|---|---|---|---|
| Open ASR English long-form WER | 10.2% ✎ | ▁▁▅▁·▃▅█▆▁▂▂ beats 10/30 | 13% |
other tests6 not in a task · 85% of its score
| test | result | vs the others in its run | counts |
|---|---|---|---|
| AfriSpeech-MultiBench · 7-region African-accent mean WER | 23.4% ✎ | ▂····▂···▂▆█ beats 6/11 | 25% |
| AfriSpeech-MultiBench · Parliament (legal domain) WER | 27.1% ✎ | ▂▂····▂·▅·▃█ beats 4/12 | 21% |
| LibriSpeech test-clean WER | 1.4% ✎ | ▁···▁▁▂·▄▄██ beats 50/56 | 17% |
| Open ASR English short-form WER (March 2026) | 7.0% ✎ | ▁▁▁▃··▁▁▂██▆ beats 61/86 | 13% |
| FLEURS English test WER | 4.2% ✎ | ▁··▁▁▁▁▁▁▃█▁ beats 45/56 | 6.8% |
| Speech Accessibility Project dysarthric speech · zero-shot WER (rated subset, N=11,218) | 15.8% ✎ | ▂·▂··▂····▆█ beats 3/9 | 1.5% |
$ where to run it
Other ways to run it
- Run it as published · needs ~5.78 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-tdt-1.1b').transcribe(['audio.wav']))"
Official weights and documentation: huggingface.co/nvidia ↗
$ how the score was worked out · AfriSpeech-MultiBench (arXiv:… 46%, Open ASR Leaderboard 27%, transcribe.cpp model catalog… 24% and 1 more
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 |
|---|---|---|
 AfriSpeech-MultiBench (arXiv:2511.14255v1) · African-accented English ASR benchmark suite | 2 | 46% |
 Open ASR Leaderboard | 2 | 27% |
 transcribe.cpp model catalog · LibriSpeech test-clean WER of validated full-precision ports | 2 | 24% |
| Dysarthric ASR with clinical context on the Speech Accessibility Project (arXiv:2605.02782v2) | 1 | 2% |
| prior · pull toward the middle | 1% | |
| weighted average | = 58.6 |
$ 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