Xiaomi's MiMo‑V2‑Pro reportedly topped OpenRouter charts after anonymous release
Rise on OpenRouter
It has been reported that Xiaomi (小米) anonymously published its new large model MiMo‑V2‑Pro on OpenRouter last week and it quickly climbed to the top of the platform’s daily charts, before becoming the weekly number one, according to a Weibo post by CEO Lei Jun (雷军). OpenRouter, which aggregates APIs for large language and multimodal models, is widely used by AI developers to pick models based on capability, latency and cost—so heavy usage is often a rough proxy for developer endorsement.
Model claims and rollout
It has been reported that MiMo‑V2‑Pro is being positioned as a flagship “agent‑era” base model: a trillion‑parameter MoE design with 42 billion activated parameters, a hybrid attention architecture and a claimed context window of up to one million tokens. Xiaomi says the model is tuned for complex workflow orchestration, long‑horizon planning and precise tool calls, with strengths in coding and automated agents; the company also announced two sibling models (Omni and a TTS system). These are being made available across Xiaomi services—MiMo Studio, Xiaomi miclaw, Kingsoft Office and the Xiaomi Browser—and accessible via developer tools such as OpenClaw, OpenCode, KiloCode, Blackbox and Cline for a limited free trial, it has been reported.
Rankings, investment and strategy
It has been reported that on the Artificial Analysis leaderboard MiMo‑V2‑Pro ranks eighth globally for overall intelligence and Xiaomi ranks fifth by brand, overtaking xAI’s Grok. Lei Jun emphasized that Xiaomi has remained “relatively low‑key” in AI but is investing heavily: the company plans to put more than 16 billion yuan into R&D and capital for AI this year. Whether that quiet rollout or aggressive funding will translate to sustained global traction remains to be seen.
Geopolitical and industrial context
Why does this matter beyond bragging rights? China’s AI firms are racing to close gaps with Western incumbents, even as U.S. export controls on advanced chips and tooling complicate access to the highest‑end training hardware. That creates a dual challenge: rapid model iteration and strong developer adoption, but potential supply‑chain and scaling constraints when serving international customers. Can Xiaomi translate early OpenRouter momentum into durable, cross‑border enterprise adoption? The next moves—hardware partnerships, benchmarks and real‑world deployments—will tell.
