TTS cost per million characters: ElevenLabs v4 promo, MAI and Sume
Per million characters: ElevenLabs v4 is $22 on its promo until Oct 12 ($80 list), MAI-Voice-2.1 is a reported $22, and Sume TTS is $47.50.
Per million characters: ElevenLabs v4 costs $22 during its promotion (72% off, until October 12) and $80 at list, MAI-Voice-2.1 is reported at $22 and its Flash model at $15, and Sume's TTS is $47.50 at $0.0475 per 1,000 characters.
Per 1M and per 100k characters
| Model | Per 1K characters | Per 100k characters | Per 1M characters |
|---|---|---|---|
| ElevenLabs v4 (promo to Oct 12) | $0.022 | $2.20 | $22.00 |
| ElevenLabs v4 (list) | $0.08 | $8.00 | $80.00 |
| ElevenLabs v4 Turbo (promo) | $0.011 | $1.10 | $11.00 |
| MAI-Voice-2.1 (reported) | $0.022 | $2.20 | $22.00 |
| MAI-Voice-2.1-Flash (reported) | $0.015 | $1.50 | $15.00 |
| Sume TTS | $0.0475 | $4.75 | $47.50 |
Reading it
The MAI numbers come from Unite.ai's coverage and Microsoft's Azure pricing page was not read, so verify them before budgeting. The promotion ends October 12, so a long project should be priced at the list rate.
An audiobook of 500,000 characters is $11.00 at the ElevenLabs promo, $40.00 at list, and $23.75 on Sume.
Not in the price
- Sume sends a plain transcript, no SSML; MAI uses SSML styles.
- Sume's maximum is 20,000 characters per request, so split long scripts.
- Voices, languages and licensing differ.
Splitting long text
Cut at paragraph breaks, keep each request under 20,000 characters and give each its own Idempotency-Key. Join the pieces with Timeline audio concat at $0.01 per job.
Sources
Related posts
More in Pricing
- Three Sume video models tie at 6.25 cents a second at 480p
Wan 3.0, MiniMax H3 and MiniMax H3 Max all bill $0.0625 per second at 480p on Sume; they split at 720p and above, so price the resolution you ship.
- Transcribe a 20-second voice note for a third of a cent on Sume
A 20-second clip costs $0.003334 on Sume STT, prorated per second at $0.01 a minute, and about $0.003 on MAI-Transcribe's $0.54 intro hourly price.
- Tripo API: $0.10-$0.30 a model, or a spin video?
Tripo's API is listed at $0.01 per credit with $0.10-$0.30 per model and GLB output. Sume has no 3D route; a 360-style product video is the alternative.
- TTS monthly bill at 1M, 5M and 20M characters: MAI vs Sume
At 1M, 5M and 20M characters a month, MAI-Voice-2.1-Flash lists $15, $75 and $300; MAI-Voice-2.1 $22, $110 and $440; Sume TTS $47.50, $237.50 and $950.
Written by Sume