Xiaomi has launched the MiMo 2.5 Pro AI model, boasting enhanced multimodal capabilities that integrate video processing, vision, and hearing. This new model, along with the MiMo 2.5, marks a significant upgrade from its predecessors by offering faster processing speeds and comprehensive modality support. The MiMo 2.5 Pro is designed for high-performance tasks, capable of handling over 1,000 tool calls independently, while the MiMo 2.5 caters to everyday users with its affordability and support for images, audio, and video. The MiMo 2.5 Pro achieves near state-of-the-art performance on coding benchmarks, solving 57.2% of tasks on SWE-bench Pro, and demonstrates token efficiency by using significantly fewer tokens than competitors. Xiaomi's commitment to AI development is underscored by its $8.7 billion investment plan, with the MiMo series playing a crucial role in maintaining its competitive edge in the AI market. The models are accessible via the Xiaomi MiMo API, with updated pricing plans and a one-time point reset bonus for existing users.