Fact · Speech recognition
NVIDIA describes Parakeet TDT 0.6B v2 as a 600-million-parameter speech recognition model designed for English transcription. Its multilingual successor is v3.
Checked against its source on October 17, 2026
In the source's own words
“parakeet-tdt-0.6b-v2 is a 600-million-parameter automatic speech recognition (ASR) model designed for high-quality English transcription”
parakeet-tdt-0.6b-v2 model card, NVIDIA, Hugging Face, 2025, Description.
Open the original- What it measures
- The card's description, with a banner pointing readers to the v3 model for 25 European languages.
- What it does not show
- A v2 model cannot transcribe the other languages. The English accuracy is the model's own strength, so comparisons with multilingual models should use English test sets.