Xiaomi launched three large models on Thursday. It officially revealed that the previously mentioned Hunter Alpha's anonymous model is the flagship base model MiMo-V2-Pro. The model has a total parameter count of 1 trillion, with 42 billion active parameters, supporting a context of 1 million tokens, ranking 8th in the AI Comprehensive Intelligence Index. It can complete complex workflow orchestration, long-range planning, and precise tool invocation without human intervention, with API pricing only 1/5 of Claude Opus 4.6. The second model, the all-modal Agent model MiMo-V2-Omni, supports cross-modal understanding of complex environments, autonomously formulates and executes plans, and can real-time adjust strategies when encountering anomalies, ultimately delivering complete results end-to-end. In the official use case provided, it utilizes OpenClaw to make purchasing decisions while browsing the web, requesting discounts on JD.com and placing orders. The third model, the speech synthesis large model Xiaomi MiMo-V2-TTS, supports multiple dialects, roles, and tones, and can intelligently recognize various format signals in the text, such as punctuation, modal particles, and emphasis markers. MiMo-V2-Pro and MiMo-V2-Omni will collaborate with the intelligent agent development framework teams of OpenClaw, OpenCode, KiloCode, Blackbox, and Cline to provide global developers with a week of limited-time free interface support.