Hundred-Million-Hour Scale
The report describes MiMo-V2-TTS as pretrained on hundred-million-hour scale data, giving it a large base for robust speech generation.
Overview
The supplied MiMo-V2 report positions MiMo-V2-TTS as the speech synthesis member of the family. It is defined by expressive voice output, support for dialect and emotion, and suitability for customer service, branded voices, and interactive assistant experiences.
Speech-Text Joint Modeling
MiMo-V2-TTS uses a multi-codebook speech-text joint modeling approach to support more controllable and expressive generation.
8K Tokens
The model is listed with an 8K context window in the supplied matrix, aligning it with targeted speech output tasks rather than long-document reasoning.
This page is grounded in the supplied MiMo-V2 research report and focuses on MiMo-V2-TTS as the Xiaomi product entity for text-to-speech and expressive voice generation.
Voice Control and Expressiveness
The report highlights MiMo-V2-TTS as more than plain speech output. It is positioned as a controllable expressive voice model.
Dialect Support
MiMo-V2-TTS is described as supporting dialect control, making it useful for region-aware voice experiences and broader accessibility across language variation.
Emotion Control
The model supports emotional delivery, allowing voice output to align more naturally with service tone, brand voice, and conversation context.
Singing Support
The supplied material explicitly includes singing among the supported capabilities, distinguishing MiMo-V2-TTS from narrower utilitarian TTS systems.
Real-Time Audio Generation
MiMo-V2-TTS is positioned for real-time audio generation, making it relevant for live assistant and voice interaction experiences.
Use Cases
The report describes MiMo-V2-TTS as a voice layer for service, branding, and more natural spoken interaction.
Customer Service Voices
MiMo-V2-TTS is suitable for customer support and assistant scenarios where voice tone and emotional nuance can improve the experience.
Branded Personas
The report highlights role-play and customized brand spokesperson use cases, where a distinct voice identity matters.
Natural Interaction
It supports paralinguistic events such as laughter and sighs, which the supplied report presents as a path to more natural and human-like interaction.
API Voice Output
The MiMo platform documentation in the supplied report states that mimo-v2-tts supports direct audio stream generation through the API.
Pricing Status
| Model | Input | Output | Positioning |
|---|---|---|---|
| MiMo-V2-TTS | Free for a limited time | Free for a limited time | Speech synthesis is temporarily not billed in the supplied report. |
Official Resources
Primary Links
Navigation
Return to the MiMo-V2 family overview to compare MiMo-V2-TTS with MiMo-V2-Pro, MiMo-V2-Flash, and MiMo-V2-Omni.