Model guide
Parakeet TDT 0.6B v3 for dictation on Mac
NVIDIA's Parakeet TDT 0.6B v3, a speech model for 25 European languages. On a Mac, Codictate runs a Core ML conversion on the Apple Neural Engine. It is also the default engine in Codictate for iPhone.
Measured in Codictate on an Apple M4 Max, at least 902 recordings per language
- Word accuracy
- 92.0%
- Wait per second of audio
- 11 ms
- Peak RAM
- 80 MB
- Download
- 500 MB
- Languages
- 25
- Translates to English
- No
Parakeet TDT 0.6B v3 accuracy by language
The share of words transcribed correctly, so higher is better. English comes from LibriSpeech audiobooks, the other languages from FLEURS read speech.
| Language | Word accuracy |
|---|---|
| English (clean) | 96.2% |
| English (noisy) | 94.1% |
| Spanish | 95.1% |
| Danish | 80.4% |
| Hungarian | 81.6% |
Word accuracy above is pooled over English (clean), English (noisy), Spanish, Danish, Hungarian: total word errors over total words spoken, not an average of the rows.
The Parakeet TDT 0.6B v3 model file
The file is parakeet-tdt-0.6b-v3-coreml (a Core ML model folder), published at FluidInference/parakeet-tdt-0.6b-v3-coreml on Hugging Face. You do not have to fetch it yourself: pick the model in Codictate's Settings and the app downloads it for you.
How it compares
Every model with a guide, from the same recordings. Wait is milliseconds per second of audio, so lower is better.
| Model | Word accuracy | Wait per second of audio | Peak RAM | Download | Languages |
|---|---|---|---|---|---|
| Whisper Large V3 Turbo Q5_0 | 93.5% | 66 ms | 778 MB | 574 MB | All Whisper languages |
| Whisper Large V3 Q5_0 | 94.0% | 106 ms | 1.5 GB | 1.1 GB | All Whisper languages |
| Whisper Large V3 | 94.0% | 164 ms | 3.5 GB | 3.0 GB | All Whisper languages |
| Whisper Large V3 Turbo | 93.6% | 93 ms | 1.9 GB | 1.6 GB | All Whisper languages |
| Whisper Medium Q5_0 | 90.8% | 65 ms | 878 MB | 514 MB | All Whisper languages |
| Whisper Small Q5_1 | 80.7% | - | 383 MB | 181 MB | All Whisper languages |
| Parakeet TDT 0.6B v3 | 92.0% | 11 ms | 80 MB | 500 MB | 25 |
| Hviske v5 Tiny Q5_0 | 88.7% (Danish) | 18 ms | 327 MB | 181 MB | Danish |
| Edda v0.2 Q5_0 | 92.5% (Danish) | 53 ms | 787 MB | 547 MB | Danish |
Which model to pick · All benchmarks, including Wispr Flow · How we measure · Parakeet TDT 0.6B v3 benchmark (Sep 2026)
Dictate with Parakeet TDT 0.6B v3
Codictate is free and open source for Mac and Windows. Press the shortcut, speak, and the text lands wherever your cursor is. The model runs on your machine, so your audio never leaves it.
Get Codictate