MiMo 2.5 TTS is Xiaomi's text-to-speech model, and on Raywake it is free to use: each request costs $0. You can pick a preset voice, describe a new voice in words, or clone a voice from your own recording, then direct the delivery with tags such as (whispering) and [pause]. It returns MP3 or WAV. Xiaomi may start charging later, and the price on Raywake would change with it.
MiMo 2.5 TTS has been live on Raywake since 30 October 2026 as MiMo 2.5 TTS. The details below come from its request contract on 31 October 2026.
MiMo 2.5 TTS at a glance
| MiMo 2.5 TTS on Raywake | |
|---|---|
| Price | Free ($0 per request) |
| Input | Text, plus optional delivery instructions |
| Preset voices | Mia, Chloe, Milo, Dean, and four Chinese voices |
| Custom voice | Describe it in words, or clone it from one audio recording |
| Delivery control | Tags such as (whispering) and [pause], and free-text instructions |
| Output | MP3 (default) or WAV |
Is MiMo 2.5 TTS really free?
Yes, for now. Raywake charges $0 per MiMo 2.5 TTS request while Xiaomi offers the model at no cost. You still need a Raywake account and, for the API, an API key. If Xiaomi starts charging, the model page will show the new price before you generate, and nothing is charged without a quote you approve.
Three ways to choose a voice
Preset voice. Set voice to Mia, Chloe, Milo or Dean for English, or to one of the four Chinese voices.
Describe a voice. Write voiceDescription, for example "a calm, warm male voice in his forties, slow pace". The model designs a voice that fits.
Clone a voice. Add one recording as inputAudio, and the speech follows that speaker's voice. Only clone your own voice or a voice you have permission to use; impersonating people is not allowed under the acceptable use policy.
How do you control the delivery?
Put direction inside the text and the model performs it:
(whispering) I have a secret. [pause] Do you want to hear it?
Use instructions for the overall reading, such as "read like an excited sports commentator" or "slow, clear narration for a training video".
When should you pick a paid voice instead?
MiMo 2.5 TTS is a strong free option for narration, video voice-overs, game characters and prototypes. Choose a paid model when you need a specific language set, a long-term fixed price, or a voice you already use elsewhere:
- MiniMax Speech-02 HD for many languages and voice settings
- ElevenLabs Multilingual v2 for natural multilingual narration
- ElevenLabs Sound Effects when you need sounds rather than speech
Our text-to-speech API comparison lists their controls and per-character prices side by side.
How to call MiMo 2.5 TTS through the API
Quote first; for MiMo 2.5 TTS the quote is $0. The model ID is mimo-2-5-tts.
curl --fail-with-body https://api.raywake.com/v1/quotes \
-H "Authorization: Bearer $RAYWAKE_API_KEY" \
-H "Content-Type: application/json" \
-d '{"model":"mimo-2-5-tts","input":{"prompt":"(whispering) I have a secret. [pause] Do you want to hear it?","voice":"Mia","outputFormat":"mp3"}}'The response has a quote_id, the price in credits (100 credits = $1) and an expiry time. See the quotes reference for the full flow.
Common questions
Is there a free text-to-speech API?
Yes. MiMo 2.5 TTS on Raywake is free per request through the same API as the paid models. You need an account and an API key.
Can MiMo 2.5 TTS clone my voice?
Yes. Add one recording of your voice as the audio input, and the speech follows that voice. Use only voices you have the right to use.
Which languages does MiMo 2.5 TTS speak?
Its preset voices cover English and Chinese. For other languages, compare MiniMax Speech-02 HD or ElevenLabs Multilingual v2.
Can I use MiMo 2.5 TTS in a video?
Yes. Generate the voice-over here, then pair it with a video model, or use it as the speech for a talking avatar such as P-Video 2 Pro Avatar.
Make speech with MiMo 2.5 TTS