Xiaomi Unveils MiMo AI Model, Shares Surge 5% on Breakthrough Launch

Xiaomi Corp (HK:1810) launched its first open-source large language model, MiMo, on Wednesday, marking its official entry into the competitive artificial intelligence space. The debut sent Xiaomi’s Hong Kong-listed shares up nearly 5% to HK$50.0, offering a modest boost to the Hang Seng Index.

MiMo is engineered for advanced reasoning tasks, positioning it as a direct rival to other cutting-edge models like DeepSeek’s R1. Xiaomi claims MiMo outperformed OpenAI’s o1-mini and Alibaba’s QwQ-32B-Preview in mathematical reasoning and coding benchmarks. The company emphasized that, using the same reinforcement learning data as DeepSeek-R1, MiMo demonstrated superior capabilities in mathematics and programming.

The move underscores Xiaomi’s push to diversify beyond its traditional smartphones and consumer electronics business. Alongside its expansion into electric vehicles with the SU7 lineup, Xiaomi is now aiming to capture a slice of the fast-growing AI market. While it already operates AI services under the Xiaomi HyperAI brand, those are currently powered by Google’s Gemini AI through an existing partnership.

The launch of MiMo places Xiaomi among a growing list of Chinese tech giants—including Alibaba, ByteDance, and DeepSeek—competing for dominance in domestic AI development. The surge in innovation follows DeepSeek’s earlier release of its R1 model, which sparked a new wave of AI research in China, despite ongoing U.S. export restrictions on advanced chips.

With MiMo, Xiaomi signals its intent to become a serious player in the AI space, leveraging open-source innovation to gain traction. The company’s strategic move aligns with a broader shift among Chinese firms toward AI-driven growth, amid rising global demand for intelligent computing tools.