MiniMax speech-2.8-hd price per 1k characters vs Sume TTS
MiniMax lists speech-2.8-hd at $0.10 per 1,000 characters and turbo at $0.06. Sume TTS bills about $0.0475 per 1,000 and takes 20,000 characters per request.
MiniMax's pay-as-you-go page lists speech-2.8-hd at $0.10 per 1,000 characters and speech-2.8-turbo at $0.06 per 1,000 (read 2026-10-03). Sume's TTS router bills about $0.0475 per 1,000 characters, which is lower than both, and accepts up to 20,000 characters per request against MiniMax's limit of under 10,000.
Price and request limits
The MiniMax T2A HTTP reference caps a single request below 10,000 characters and exposes speed from 0.5 to 2 and pitch within plus or minus 12. Sume's TTS router exposes generation_config with speed from 0.6 to 1.5, volume from 0.5 to 2 and an emotion string.
| Item | MiniMax speech-2.8-hd | MiniMax speech-2.8-turbo | Sume TTS router |
|---|---|---|---|
| Price per 1,000 characters | $0.10 | $0.06 | $0.0475 |
| Max characters per request | Under 10,000 | Under 10,000 | 20,000 |
| Speed range | 0.5 to 2 | 0.5 to 2 | 0.6 to 1.5 |
A 5,000-character script
MiniMax hd: 5 x $0.10 = $0.50. MiniMax turbo: 5 x $0.06 = $0.30. Sume: 5 x $0.0475 = $0.2375, rounded up to $0.24 per job. Price is only one axis; this post does not rank voice quality, which you should audition on your own script.
Voice cloning cost
MiniMax lists rapid voice cloning at $1.5 per voice, charged on the first synthesis. Sume's TTS router selects a voice with avatar_id, avatar_handle or voice.id, and rejects a malformed id with invalid_voice_id. This post makes no cost claim for creating a clone on Sume, because that price is not part of the TTS router catalog.
Which to choose
Pick MiniMax when its voices or languages fit your project and you stay under its request cap. Pick Sume when the speech needs to land inside a Sume render or caption job. See the 1,200-second limit post for the duration ceiling on Sume.
Sources
Related posts
More in Pricing
- Balance needed for one AI generation call: published max per endpoint
Sume's catalog publishes estimated, minimum and maximum cents per endpoint. The worst-case hold for one call runs from 3 cents to $56.25 before a 402 can fire.
- Monthly AI video budget: 60 five-second clips per month on Sume
Sixty 5 second clips a month cost $30.00 on H3 Max 768p, $37.80 on Wan 3.0 or Gemini Omni Flash 720p, $113.40 on Seedance 2 and $173.40 on Seedance 2.5.
- Monthly content calendar budget: 12 Reels, 4 Shorts, 8 carousels
A month of 12 Reels, 4 Shorts and 8 carousels runs from about $83 to $372 on Sume by model tier. Per-post finishing costs and retake allowance shown.
- Most expensive single AI video request on Sume: max cost per model
The ceiling for one request on Sume: Genjutsu at 30 s costs $25.54, Recast at 30 s 1080p $16.88, Wan 3.0 at 30 s 1080p $7.50. Table for every priced row.
Written by Sume