How many 600-character Shorts voiceovers does $5 buy? MAI vs Sume

At 600 characters a script, $5 buys about 555 Shorts voiceovers on MAI-Voice-2.1-Flash ($15 per 1M characters), 378 on MAI-Voice-2.1 ($22 per 1M) and 175 on Sume TTS 1.0 ($47.50 per 1M, or $0.0475 per 1,000 characters).

Rates are from Microsoft's announcement and the Sume API reference and pricing page, read 2026-10-05. The 600-character script is an assumption for a short vertical video; count your own.

The conversion

One script is $0.009, $0.0132 or $0.0285. The number of scripts per $5 is the budget divided by that, rounded down.

Reading the gap

Sume costs about 3.2 times MAI-Voice-2.1-Flash per character and about 2.2 times MAI-Voice-2.1. The price buys you a job-based API where the audio, word timings and sentence slices come back as hosted files.

• Sume TTS 1.0 accepts up to 20,000 characters per request, so a 600-character script is far below any ceiling.
• Add timestamps.words to get word timings, which a caption render needs.
• Voices are chosen by avatar_handle, avatar_id or voice.id; the language is a separate language field.
• Microsoft's Flash announcement lists 45 seconds of audio generation, which suits a short script.

Budget check in Python

This prints how many scripts a budget buys. Change the script length or the budget to match your own.

import math
budget, chars = 5.0, 600
rates = {"MAI-Voice-2.1-Flash": 15.0, "MAI-Voice-2.1": 22.0, "Sume TTS 1.0": 47.5}
for name, per_million in rates.items():
    cost = chars * per_million / 1_000_000
    print(f"{name}: {cost:.4f} USD each, {math.floor(budget / cost)} scripts")