Get Zlio on your home screenOne tap, full screen, like an app.
AI Voice · first seen 29 October 2026, 07:09 UTC
Nvidia releases Nemotron 3.5 Transcribe open speech model
An open-weight ASR model claiming rival accuracy at far lower active compute.
Nvidia released Nemotron 3.5 Transcribe, a free open-weight speech-to-text model with 30B parameters covering 40 languages. Nvidia says it matches rivals using one-sixth the active parameters and cut errors 62% on Indian doctors' dictation.
