Five takes of an 800-character script: MAI Flash, MAI 2.1, Sume TTS

Re-recording an 800-character script five times to pick the best read costs $0.06 on MAI-Voice-2.1-Flash, $0.088 on MAI-Voice-2.1 and $0.19 on Sume TTS. A single take is $0.012, $0.0176 and $0.038 respectively.

Why price the takes

A script is rarely read once. At these rates even 20 takes cost under a dollar on all three, so the choice should rest on voice quality, not the audition bill. Microsoft's prices are on its announcement page.

Sume side

Sume TTS is $0.0475 per 1,000 characters, prorated, per the rate card. Each request takes up to 20,000 characters, so you can also audition several short lines in one request.

Recompute it in Python

Characters times takes times price. The numbers are the ones in the table above.

chars, takes = 800, 5
for name, per_million in (('MAI Flash', 15), ('MAI 2.1', 22), ('Sume TTS', 47.5)):
    print(name, round(chars * takes * per_million / 1e6, 4))