Xiaomi MiMo-V2.5 Series Models Brand New Release
Token Plan now supports monthly/annual subscriptions, fully covering V2.5 models.
One account is eligible for one first purchase discount.
Models Introduction
Token Plan Available
mimo-v2.5-pro
Flagship foundation model. Peak intelligence.
- Trillion Parameters, Efficient Architecture — 1T total params | 42B active | 1M extended context
- Ultimate Agent Performance — Matches Claude Opus 4.6 in demanding agentic scenarios
Token Plan Available
mimo-v2.5
Next-Level Omni-Modal Agent Foundation
- Native Omni-Modal Perception + 1M Context — Natively understands images, video, audio, and text, enabling precise cross-modal perception and long-range reasoning — among the best in the industry.
- Powerful Omni-Modal Agent — Built-in agentic capabilities for browsing, understanding, reasoning, and execution of complex tasks, rivaling mimo-v2.5-pro on everyday tasks.
- Performance Meets Efficiency — Achieves superior token efficiency while maintaining leading capabilities, sitting on the Pareto frontier of performance and cost.
Token Plan Available
MiMo-V2.5-TTS Series
Voice Synthesis Models Built for the Agent Era
- Premium TTS Voices — Built-in high-quality voices with strong style-instruction following — fine-grained control over speed, emotion, and tone for any scenario.
- Voice Design — Define and generate entirely new voices from a single text prompt — fast, intuitive, and creative.
- Voice Cloning — High-fidelity voice replication from minimal audio samples, with consistent timbre, strong generalization, and robust stability.
Invite friends, get API credit
Invite friends to MiMo — you both get $2 in API credits
Instant on signup · Credits valid 40 days