Voice AI in vogue: Nvidia releases new open transcription model

modelsTechCrunch AIOct 28, 2026

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 ElevenL...

Impact: LowConfidence: MediumNovelty: Incremental
modelstechcrunch-aimultimodal

See the whole story

Read the original on TechCrunch AI

Sens.ai aggregates and assesses published reporting. The assessment above is machine generated; the original source is authoritative.