TTS price per 1,000 characters: every Sume row next to MAI-Voice

Per 1,000 characters, Sume's TTS Router bills $0.0475 on every Sonic row, $0.05 on Eleven v4 Turbo, $0.0625 on Eleven Turbo v2.5, $0.10 on Eleven v4 and $0.125 on Eleven v3 and Multilingual v2, all before rounding the job total up to the next cent. Microsoft lists MAI-Voice-2.1 at $22 per million characters ($0.022 per 1,000) and Flash at $15 ($0.015).

Sume values are the catalog's provider list per character times the 1.25 house margin, from packages/api-contract/src/tts-router-models.ts (origin/main b28f3b0). Microsoft's prices are from its model page. Both read on 2026-10-13. The Sume figures include Sume's margin and the vendor figures do not, so this is a rate comparison, not a like-for-like bill.

What does a 15 second ad line cost?

A 15 second line is about 200 to 250 characters at a normal pace. At $0.0475 per 1,000 characters that is about $0.012 of characters, but the catalog's minimum is one cent, so expect one cent. The cent minimum is why per-line prices look flat for short lines and diverge only past a few thousand characters.

Where do I read the live price?

The router catalog publishes list_usd_micros_per_character, billable_margin and the formula for each row, so a script can compute a quote before it submits. Admission preview and MCP dry_run also return an estimate before any credit is reserved.

curl https://api.sume.com/v1/tts-router/models \
  -H "Authorization: Bearer $SUME_API_KEY"