Skip to content

Nova-3

Built for streaming rather than batch, with the low latency and concurrency handling that live captioning and voice agents need. Deepgram also sells a managed Voice Agent API on top, which bundles the LLM and TTS legs and prices per conversation minute rather than per transcription minute.

Pricing checked against provider documentation on . How we verify

stable Deepgram • Nova • nova-3

Price

$0.0043

/ audio min

Context

—

not token-based

Max output

—

varies

Pricing caveat: Batch pay-as-you-go rate for the monolingual model. Streaming is higher and quoted between $0.0048 and $0.0077/min depending on plan and source; multilingual adds roughly 20%. Confirm your tier in the console.

Where Nova-3 fits

  • Live captioning
  • Voice agents and IVR
  • Real-time meeting transcription
  • Call-centre analytics

Specifications

API model ID
nova-3
Provider
Deepgram
Model family
Nova
Released
February 2025
Knowledge cutoff
—
Context window
—
Max output
—
Native reasoning
No
Relative latency
low
Input modalities
audio
Output modalities
text
Open weights
No
Cached input
—
Batch discount
—

Frequently asked

How much does Nova-3 cost?

Nova-3 is billed at $0.0043 / audio min.

What is Nova-3's context window?

Context limits for Nova-3 depend on the input format rather than a token count.

Is Nova-3 a reasoning model?

No. Nova-3 answers directly without a native thinking phase, which keeps latency and cost predictable.

What is Nova-3 best for?

Real-time streaming transcription. Built for streaming rather than batch, with the low latency and concurrency handling that live captioning and voice agents need. Deepgram also sells a managed Voice Agent API on top, which bundles the LLM and TTS legs and prices per conversation minute rather than per transcription minute.