Nvidia Unveils Open-Source Speech-to-Text Model for Multilingual Transcription

Nvidia released Nemotron 3.5 Transcribe, a free open-source speech-to-text model supporting 40 languages for recorded audio, targeting business and medical transcription applications. The 30-billion-parameter model emphasizes multilingual accuracy and hardware efficiency but excludes real-time processing capabilities.

Nvidia just released a free speech-to-text model that can transcribe up to 40 languages, but it only works on recordings, not live audio, so it’s a big step for meeting and medical transcription tools while still leaving real-time use to rivals like ElevenLabs.

Read original article