Question · Speech recognition
How much does Deepgram cost?
Deepgram's Nova-3 model lists at $0.258 per hour of recorded English audio on pay as you go, and live streaming costs more, from $0.462 per hour at the regular rate. New accounts get a $200 free credit, which covers about 775 hours of recorded audio.
Updated October 27, 20263 min read16 cited facts, each traced to its source
Deepgram charges by the amount of audio you send, not by seat or subscription. On pay as you go, its main model, Nova-3 Monolingual, lists at $0.258 per hour of recorded audio, which is $0.0043 per minute 1. Live (streaming) audio costs more: $0.462 per hour at the regular rate 2. Every new account starts with a $200 free credit 3.
These prices were read on Deepgram's pricing page on 27 October 2026. Deepgram lists prices per minute; the hourly figures here are the per-minute price times 60. Prices change, so confirm them before you budget.
Recorded audio costs $0.216 to $0.312 per hour
The price depends on whether you send a file or a live stream, whether the audio is in one language or several, and which plan you are on. For recorded files:
| Model | Pay as you go, per hour | Growth plan, per hour |
|---|---|---|
| Nova-3 Monolingual | $0.258 1 | $0.216 4 |
| Nova-3 Multilingual | $0.312 5 | $0.258 6 |
| Whisper Large (hosted) | $0.288 7 | $0.288 |
Multilingual audio, where speakers switch languages, costs about 21 percent more than single-language audio on pay as you go 8. For recorded files, speaker labels ( speaker diarization) and smart formatting are included in the model price.
Live audio costs more, and the cheaper rate is a promotion
Nova-3 Monolingual streaming lists at a regular $0.0077 per minute, with a current promotional rate of $0.0048 per minute 2. Over an hour, the promotion works out to about $0.288 against $0.462 regular 9. The pricing page called the streaming rates "limited-time promotional" and showed no end date on the day we read it. Budget at the regular rate and treat the promotion as a discount you may lose.
Add-ons raise a streaming bill. Speaker diarization on live audio is an extra $0.0020 per minute 10, and Keyterm Prompting, which biases the model toward your product names and jargon, adds $0.0013 per minute 11.
The free tier is a $200 balance, not a monthly allowance
The $200 credit is a one-time balance that then rolls into pay as you go 3. Spent on Nova-3 Monolingual recorded audio, it covers about 775 hours 12. Live audio or multilingual audio uses it faster. That is enough to test several hundred hours of your own audio before paying anything.
The plans differ in commitment, not in the model
- Pay as you go: no minimum, no expiry and no credit card to start 13.
- Growth: a prepaid commitment of $4,000 or more a year, which cuts rates by up to 20 percent 13.
- Enterprise: custom pricing through sales, and the tier where Deepgram signs HIPAA business associate agreements 14.
It depends: when the real cost differs from the list price
- Live audio with extras. A voice agent that needs speaker labels and keyterms pays the model rate plus both add-ons, so price the whole bundle.
- Resellers. Cloud platforms resell the same model at their own rates. Cloudflare's Workers AI lists Nova-3 at $0.312 per hour 15. If you buy that way, the reseller's terms apply as well as Deepgram's.
- Rivals. AssemblyAI's Universal-3.5 Pro lists at $0.21 per hour, so Nova-3 costs about 1.23 times as much on recorded audio 16. See Deepgram vs AssemblyAI and how much AssemblyAI costs.
What to do
- Sign up, take the $200 credit, and transcribe 30 to 60 minutes of your own audio with Nova-3.
- Price your real monthly volume at the regular rates, with the add-ons you need.
- Consider Growth only if your yearly spend will clearly pass $4,000.
- If accuracy is your main worry, read how accurate Deepgram is before you decide.
For the full picture, including data retention and deployment options, see our guide to Deepgram and the wider overview of speech recognition. If you only want to dictate on your own computer, an API is the wrong tool: free browser dictation at /online-dictation or an on-device app avoids per-hour billing altogether.
Evidence
Every number on this page, with where it comes from. Open a fact to see the exact words in its source.
- 1.
- 2.
- 3.
- 4.
- 5.
- 6.
- 7.
- 8.
- 9.
- 10.
- 11.
- 12.
- 13.
- 14.
- 15.
- 16.
The data behind these numbers
Go deeper
Related questions
- Deepgram vs AssemblyAI: which speech-to-text API is better?
- How accurate is Deepgram?
- Are open-source speech recognition models as accurate as paid APIs?
- AssemblyAI vs OpenAI Whisper: which is more accurate?
- Can an ESP32 or microcontroller do speech recognition?
- Can large language models do speech recognition?
- Can Whisper AI do text to speech?
- Can Whisper identify different speakers?