Xiaomi MiMo-V2.5-TTS Series

MiMo-V2.5-TTS Series includes three models, available on Xiaomi MiMo API platform, limited time free.

Three Models

• MiMo-V2.5-TTS: Standard voice synthesis
• MiMo-V2.5-TTS-VoiceDesign: Text description → new voice
• MiMo-V2.5-TTS-VoiceClone: Voice cloning

Features

• Natural language control for emotion, speed, style
• Audio tag control
• Dialects: Sichuan, Henan, Cantonese, Taiwan accent, etc.
• Multi-Codebook Joint Modeling architecture

Notes

• TTS is API-only (not fully open source for local deployment)
• ASR is open source (Apache-2.0)
• Limited time free on platform