Speech 02
minimax/speech-02Production text-to-speech with expressive multilingual voices, priced per character.
— ctx · Price on request
Category
Speech
Text-to-speech and speech recognition. Natural voices out, accurate transcripts in, the interface layer of spoken products. Reach for speech at the edges of a product, wherever a person talks to it or listens to it. Synthesis is billed of script rather than per , so cost is easy to forecast from copy you already have.
2 models in Speech from 2 providers
R/1M chars Speech is billed by characters of script, so the cards below show a per-character price and no context window.
Top models
Leading speech models
Browse all speech models →
2 models in the directory
How to choose
Choosing a model in Speech
- 01
Character pricing
TTS bills per character, not per token. Estimate from your actual script lengths.
- 02
Voice quality vs speed
Interactive agents need low first-audio latency; audiobook rendering can favour quality.
- 03
Language coverage
Voice naturalness varies enormously by language. Audition in every locale you ship.