Fact · Speech recognition
24minutes
NVIDIA says Parakeet TDT 0.6B v3 handles audio up to 24 minutes long with full attention, or up to 3 hours with local attention.
Checked against its source on October 17, 2026
In the source's own words
“supporting audio up to 24 minutes long with full attention (on A100 80GB) or up to 3 hours with local attention”
nvidia/parakeet-tdt-0.6b-v3 model card (Hugging Face), NVIDIA, Hugging Face, 2025, Key features, long audio transcription.
Open the original- What it measures
- The limit is stated for a single A100 80GB GPU, which is the reference hardware NVIDIA names.
- What it does not show
- The limit depends on the GPU and the attention mode. A home computer with less memory will handle shorter files, so split long recordings.