Deepgram Nova-3 — reviews, specs & pricing

Enterprise speech-to-text model built for noisy, multi-speaker real-time audio.

Summary

Deepgram's most accurate speech-to-text model, tuned for noisy, multi-speaker enterprise audio. It extends the Nova line's lead in real-time transcription latency.

Sample use case

Enterprise real-time and batch speech transcription.

Specifications

  • Provider: deepgram
  • License: closed
  • Released: 2025-02-12

Average rating 0.0 from 0 community reviews on Reviuws.

Frequently asked questions

What is Deepgram Nova-3?

Deepgram's most accurate speech-to-text model, tuned for noisy, multi-speaker enterprise audio. It extends the Nova line's lead in real-time transcription latency.

How much does Deepgram Nova-3 cost?

Pricing for Deepgram Nova-3 is not publicly listed.

Is Deepgram Nova-3 open source?

Deepgram Nova-3 is released under the closed license.