Xiaomi Releases MiMo-V2.5, Enhancing Agentic and Multimodal Capabilities
Xiaomi has released MiMo-V2.5, a 310-billion-parameter Sparse Mixture of Experts model with 15 billion active parameters. This model significantly improves agentic capability and multimodal understanding, offering native visual and audio understanding and supporting up to 1 million tokens of context. It achieves frontier-level agentic performance with high token efficiency and is now open-sourced.
This release brings advanced multimodal and agentic AI capabilities to an open-source model, making sophisticated AI more accessible for development and deployment. Its high token efficiency at a competitive performance level could reduce operational costs for AI applications.
Learn one new AI thing every day.
Daily Deck sends you seven plain-English cards like this every morning. Free.
Start free