New AssemblyAI-TTS-1 model targets low-latency voice agent performance
In short: AssemblyAI has released AssemblyAI-TTS-1, a text-to-speech model built for voice agents. It features sub-100ms latency, 4,000+ available voices, and instant voice cloning, priced at $15 per 1M characters without concurrency limits.
This summary was generated automatically by AI from AssemblyAI's publication. It is our own text, not a copy of the original — facts, figures and quotes belong to the source, linked above and below.
What changed?
- 1Median time to first audio of 47 ms.
- 2Pricing set at $15 per 1M characters.
- 3Unlimited concurrency included with pay-as-you-go pricing.
- 4Support for English, Spanish, French, German, Italian, and Portuguese.
| Parameter | Before | Now |
|---|---|---|
| Time to first audio (median) | Not specified | 47 ms |
| Price | Not specified | $15 per 1M characters |
Why it matters
For contact-center automation, latency is a primary driver of caller frustration. This model provides an alternative that focuses on natural, fast responses to improve call completion rates.
What it means for AI agents and contact centers
When testing this against current providers, measure the time-to-first-byte and entity accuracy (dates, numbers, codes). Compare the integration overhead for your current LiveKit or Pipecat pipelines as this model supports native streaming and barge-in.
🧪 Worth Testing
Testing is recommended to see if the 47ms latency significantly improves user engagement compared to your existing TTS providers.
Sources
- AssemblyAIOfficialPrimary sourceOriginal article →„Introducing AssemblyAI-TTS-1: text-to-speech built for voice agents“
- Published by source
- —
- Found by our system
- 28 Oct 2026, 14:52
- Summary generated
- 28 Oct 2026, 15:22
This article was written by AI from the original source. Facts, numbers and prices come from the source; missing values are marked “Not specified”. Legal notice, copyright and privacy