MiMo V2.5 TTS
The standard V2.5 TTS model turns text into speech using Xiaomi's built-in voices. It supports style control and a singing mode.
The spoken text and direction are separate
The release supports natural-language direction and inline audio tags. The standard preset-voice model supports singing; its VoiceDesign and VoiceClone siblings do not.MiMo V2.5 TTS family announcement
Choose a stable voice for repeat content
Practical guidance. Preset voices are useful when many short messages must sound consistent. Review names and numbers in the intended language, and keep the style prompt short enough that it does not compete with the words being spoken.
Specifications & access
| Detail | Value and conditions |
|---|---|
| Access | API |
| Model identifiers | mimo-v2.5-tts |
| Input arrangement | Put the words to be spoken in an assistant message. A user message can carry style instructions.MiMo V2.5 speech synthesis guide |
| Streaming | The standard model supports streaming speech; the guide requires PCM16 for streaming output.MiMo V2.5 speech synthesis guide |
| Model boundary | Voice design and voice cloning use separate model IDs.MiMo current model catalogue |
Related models
- MiMo V2.5 ASR
Xiaomi · MiMo V2.5 Speech
V2.5 ASR targets transcription across Mandarin, English, Chinese dialects, mixed-language speech and noisy recordings.
Launch date not confirmedAPI · Downloadable - MiMo V2.5 TTS VoiceClone
Xiaomi · MiMo V2.5 Speech
VoiceClone generates speech using a voice sampled from audio.
Launch date not confirmedAPI - MiMo V2.5 TTS VoiceDesign
Xiaomi · MiMo V2.5 Speech
VoiceDesign creates a new voice from a written description.
Launch date not confirmedAPI