One 280-character voice line: MAI Flash vs MAI 2.1 vs Sume TTS

A 280-character voice line, one tweet's worth, costs about $0.0042 on MAI-Voice-2.1-Flash, $0.0062 on MAI-Voice-2.1 and $0.0133 on Sume TTS. All three are far below a cent per line to a couple of cents, so the difference shows up only when you scale to thousands of lines.

Method

Characters times price per million for Microsoft, from the announcement; characters times $0.0475 per 1,000 for Sume, from the rate card. Sume's per-request charge is prorated in micro-dollars, so one 280-character line is $0.0133.

When one line is not one request

Each Sume TTS request takes up to 20,000 characters, so batch short lines into one script when your use case allows it. Batching does not change the per-character price, but it does cut round trips.

Recompute it in Python

Sume's $0.0475 per 1,000 characters is $47.50 per million, which puts all three on the same scale. The numbers are the ones in the table above.

chars = 280
for name, per_million in (('MAI Flash', 15), ('MAI 2.1', 22), ('Sume TTS', 47.5)):
    print(name, round(chars * per_million / 1e6, 4))