Xiaomi has officially announced a groundbreaking integration for its AI ecosystem. The highly advanced Xiaomi MiMo models are now fully integrated with Hermes Agent, one of the world’s most watched open-source agent frameworks. To celebrate this milestone and encourage global innovation, Xiaomi is offering a comprehensive two-week free trial for developers worldwide.
New Features
This update brings a massive leap in AI agent capabilities by fusing Xiaomi’s state-of-the-art language models with the Hermes Agent framework. Hermes Agent is renowned for its ability to self-evolve, maintain cross-session memory, and automatically accumulate task experience—becoming smarter and more capable with every interaction across all supported platforms.
The integration introduces full support for three specific models, each tailored for different computational needs:
-
Xiaomi MiMo-V2-Pro: Built for complex logic and deep analysis, this model features a massive 1M long context capability. It includes native, powerful tool-calling functions and deep Agent-specific optimizations that perfectly align with Hermes’ self-evolving skills and complex workflow requirements.
-
Xiaomi MiMo-V2-Omni: This model significantly expands the boundaries of machine perception by integrating full-modal understanding. It seamlessly processes images, video, audio, and text simultaneously, transforming Hermes Agent into a truly visible, audible, and action-capable full-modal agent.
-
Xiaomi MiMo-V2-Flash: A highly optimized, lightning-fast model designed for rapid token generation and low-latency tasks.
Supported Platforms
-
Supported AI Models: Xiaomi MiMo-V2-Pro, Xiaomi MiMo-V2-Omni, Xiaomi MiMo-V2-Flash
-
Target Framework: Hermes Agent (Latest Version required)
-
Access Platform: Nous Portal
Market Impact
This rollout dramatically lowers the barrier to entry for developers building complex, full-modal AI applications. During the two-week promotional period (running from April 8 to April 22, 24:00 UTC+8), global developers can access these top-tier models completely free of charge. As verified by the Nous Portal interface, the token price for the Pro, Omni, and Flash models has been temporarily reduced to $0.00/1M tokens. By making these powerful APIs accessible, developers can stress-test cross-session memory and full-modal logic without worrying about overhead API costs.
he was a baby