MiMo-V2.5
MiMo-V2.5 is a native omnimodal model by Xiaomi. It delivers Pro-level agentic performance at roughly half the inference cost, and beats MiMo-V2-Omni on image and video understanding.
Specifications
| Lab | Xiaomi |
|---|---|
| Context window | 1.1M tokens |
| Input price | $0.14/1M |
| Output price | $0.28/1M |
| Release | Apr 2026 |
Strengths
- Native omnimodal vision-language-audio
- Pro-level agentic performance at lower cost
- 1M context window
- Strong image and video understanding
Weaknesses
- Below MiMo-V2.5-Pro on complex tasks
- Limited Western market presence
Best for
- Multimodal agent workflows
- Video and image understanding tasks
- Cost-effective omnimodal applications
See what MiMo-V2.5 built
Interactive pages generated from the same creative briefs given to every model.
Cambrian Open World
Generate an explorable 3D Cambrian seafloor populated by strange prehistoric life.
Quantum-Powered AI Lab
Generate a credible landing page for an AI research lab powered by quantum computing.
Safari Reserve Manager
Generate a 3D management game about operating a wildlife reserve and safari business.
In Depth: MiMo-V2.5
Summary
MiMo-V2.5 is an AI model from Xiaomi.
Released in Apr 2026. It currently appears in the Overall category on LMRank and 1 other category. It takes text, audio, image, and video input and produces text output, with a context window of 1.1M tokens. Input pricing is $0.14 per 1M tokens and output is $0.28 per 1M tokens on OpenRouter.