Is MAI-Voice-2.1 on Sume? What the TTS Router lists instead

Short answer

No. As of 2026-10-08 the Sume TTS Router catalog lists Sonic models only, and no Microsoft MAI voice model. Microsoft AI's news page names MAI-Voice-2.1 and MAI-Transcribe-2, but naming a model is not availability on Sume. If a post or a chat answer says Sume exposes MAI-Voice-2.1, it is wrong.

What you can call today is the TTS Router at POST /v1/tts-router/generate with a required model id from GET /v1/tts-router/models.

The catalog as of 2026-10-08

All five rows share the same limits and the same price; they differ in release channel. The list price is 38 micro-dollars per character and the billable price is list x 1.25, rounded up to cents per job.

How the vendor listings compare

A release tracker lists MAI-Voice-2.1 at $22 per million characters in 23 languages, and MAI-Voice-2.1-Flash at $15 per million with a vendor-stated 150 ms latency. Vendor prices come from a third-party release tracker that lists them, not from a Microsoft price page, so confirm them on Microsoft's own pricing before you commit.

Sume's billable rate works out to $47.50 per million characters before per-job rounding (38 x 1.25 = 47.5 micro-dollars). The comparison is price only: this post says nothing about voice quality for either side.

What to do if you need a Microsoft voice

Call it from Microsoft or another provider that carries it, then bring the finished audio into Sume for the rest of the workflow. Timeline audio joins and slices Sume-hosted audio (Timeline audio), so import the file first; the server does not fetch arbitrary URLs.

Everything downstream, captions, music and Timeline renders, works on the audio file, not on which engine made it.

How to verify this yourself

The catalog is the source of truth, not a blog post. Call GET /v1/tts-router/models with your API key and read the ids. Unknown model ids fail with 400 model_not_found and the catalog URL, so if you try a Microsoft id you will be pointed back to the list. The same applies to the music and video routers: the catalog lists what is callable.

If Sume adds a model, the catalog changes before any marketing page does. A good habit for any provider is to check the live catalog at deploy time rather than hard-coding an id from a blog.

Why name the competitor at all

Buyers search for the Microsoft model name and want to know what a switch would cost. The honest answer is a price comparison plus a clear statement of what Sume does not carry. Prices differ per character, and so do the surrounding features, such as joining, captions and rendering, which Sume offers as separate jobs. Pick based on the whole workflow, and test a short script on both before moving a production series.