Decagon Voice 3 price: what is public and how to size a pilot
Decagon's Voice 3 launch post lists no price. How to size a voice-agent pilot in call minutes, and what the speech parts cost on public rate cards.

The Voice 3 launch post and the Dialogues 2026 post that we read list no price, so the only honest answer to what Decagon Voice 3 costs is to ask Decagon for a quote. You can still size a pilot in call minutes now and compare it with the public prices of the speech pieces.
Size the pilot first
Pricing quotes depend on volume, so bring numbers.
- Calls per month, split by easy and hard.
- Average minutes per call.
- Share you expect the agent to resolve without a human.
- Languages needed, since Decagon lists 70+.
- Whether you need after-call transcripts for review.
Public prices for the speech parts
For comparison only, the building blocks of a do-it-yourself stack have public list prices. A packaged agent adds reasoning, tools, guardrails and support on top.
| Part | List price | Source |
|---|---|---|
| Streaming transcription | $0.54 per hour introductory, through the end of 2026 | Microsoft AI |
| Speech from text, Flash | $15 per 1M characters | Microsoft AI |
| Batch transcription (Sume STT) | $0.01 per audio minute | Sume rate card |
A rough example
For 6,000 minutes a month, the Microsoft streaming transcription rate comes to about $54 at $0.54 per hour (100 hours). That is one small part of the cost. The rest is the language model, tools and ops, so a quote is the only way to compare.
Use Sume for review, not for the live call
After a pilot starts, you will want recordings reviewed. Sume's STT turns a recording into text and word timings at $0.01 per audio minute, so reviewing 6,000 minutes costs about $60 at the public rate. Each job reserves up to 10 minutes, so split longer calls first.
Sources
Related posts
More in Pricing
- Does an image cost more at 16:9 than 1:1 on Sume? Flat rows vs GPT 2.5
On most Sume image rows the ratio does not change the price. ChatGPT Image 2.5 is the exception: cost follows pixels and quality. Billed figures for both.
- Draft 20 Wan 3.0 clips at 480p, then finish the best six at 1080p
Twenty 5-second Wan 3.0 drafts at 480p cost $6.40; finishing six at 1080p adds $7.50. All twenty at 1080p cost $25.00.
- Price per kept AI clip when 1 in 3 is rejected: Wan 720p
A 15-second Wan 3.0 clip at 720p is $1.875 on Sume. If you keep one clip in three, the real price per kept clip is $5.625. The keep rate is a budget input.
- 8-second 720p Seedance price ladder: mini, fast, 2.0 and 2.5 on Sume
One 8-second 720p clip costs about $1.51 on seedance-2-mini and $4.62 on seedance-2.5. The token formula behind each Sume price, with the 1.25 margin shown.
Written by Sume