4,750-character newsletter read aloud: MAI-Voice-2.1 vs Sume TTS cost

The answer for one 4,750-character issue

On Sume TTS a 4,750-character newsletter issue is one job and costs 23 cents ($0.23). At the per-million-character rates a tracker lists for Microsoft's voice models, the same text is $0.1045 on MAI-Voice-2.1 and $0.0712 on MAI-Voice-2.1-Flash. Sume does not list MAI-Voice-2.1, MAI-Voice-2.1-Flash or MAI-Transcribe-2 in its catalogs, so this is a price comparison, not a statement that you can call Microsoft's models through Sume.

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.

The arithmetic

Sume TTS list price is 38 micro-dollars (0.0038 cents) per character; the billable price is list x 1.25, rounded up to whole cents per job. That is 0.00475 cents per character before rounding, so a 20,000-character request is 20,000 x 0.00475 = 95 cents. For this issue: 4750 x 0.00475 = 22.5625 cents, rounded up to 23 cents.

The vendor side is characters divided by one million, times the listed rate.

What the gap means for a weekly send

Over 52 issues a year the Sume figure is 52 x 23 = 1196 cents ($11.96). The 2.1 rate would be $5.434 and Flash $3.705. The difference is a few dollars a year, so for a single newsletter the choice is about voice quality and workflow, not cost.

The rounding matters at the other end: very short texts pay the one-cent floor per job, so batching many short lines into one request is cheaper than one job per line, as long as you can cut the audio apart afterwards.

Where Sume fits the workflow

Sume returns the audio as a durable file you can join, caption and place on a timeline. The job record keeps the model, voice, language and speed, so next week's issue can reuse them (Jobs and results). Clips can be joined without re-synthesis through Timeline audio.

If you only need a price per character, use the table. If you need captions and a rendered video from the same voice track, count the other line items too: captions are $0.20 per job and Timeline is $0.10 per output minute.

Checks before you choose

Run the same two paragraphs through each service and listen with the same headphones. Price per issue is a few cents either way, so a voice your listeners dislike costs more than any rate difference. Check how each service handles numbers, product names and links read aloud, and rewrite those passages in the source text if needed.

Also check how the audio will be delivered. A newsletter read-aloud usually ships as an mp3 link, so confirm that the output format you request is the one your player accepts. On Sume, the job record shows the output format used, which makes it easy to keep every issue identical.

A simple monthly model

Four issues a month at 4,750 characters is 19,000 characters. Sume bills each issue as its own job, 4 x 23 = 92 cents. If you instead batch the month into a single request under the 20,000-character cap, the bill is 19,000 x 0.00475 = 90.25 cents, rounded up to 91 cents. The saving is one cent, so batch only when it suits your editorial process, not for the price.