Xiaomi Unmasked as Creator of Mystery Hunter Alpha AI Model After Week of DeepSeek Speculation
An anonymous trillion-parameter AI model that topped OpenRouter usage charts and triggered widespread speculation about a stealth DeepSeek V4 launch turned out to be Xiaomi's MiMo-V2-Pro, signaling the smartphone and EV maker's escalating ambitions in the AI model race.
The Mystery Appears
On March 11, an anonymous AI model appeared on the developer platform OpenRouter under the codename Hunter Alpha. It carried no developer attribution, and OpenRouter itself labeled it a stealth model. Its listed specifications were striking: one trillion parameters and a context window of up to one million tokens. Within days, the model had surpassed one trillion tokens in total usage and climbed to the top of OpenRouter’s leaderboard, according to The Japan Times.
The AI community quickly assumed the model was a stealth test of DeepSeek’s anticipated V4 system. The circumstantial evidence appeared strong: during testing by Reuters, the chatbot introduced itself as a Chinese AI model primarily trained in Chinese, and reported a training data cutoff of May 2025 — the same knowledge cutoff that DeepSeek’s own chatbot reported, as The Japan Times noted. Chinese media outlets had previously reported that DeepSeek V4 might arrive as early as April, making the timing of Hunter Alpha’s appearance particularly suggestive.
Xiaomi Steps Forward
On March 18, Xiaomi’s AI division MiMo confirmed that Hunter Alpha was an early internal test build of MiMo-V2-Pro, the company’s flagship large language model designed to power autonomous AI agents. The team is led by Luo Fuli, a former core contributor to DeepSeek’s breakthrough models who joined Xiaomi in late 2025, bringing significant architectural expertise from one of China’s most recognized open-source AI labs.
The revelation that an anonymous model mistaken for DeepSeek’s next-generation system was built by a company better known for smartphones and electric vehicles underscored the breadth of China’s AI development landscape. MiMo-V2-Pro features over one trillion total parameters with 42 billion active parameters, roughly three times larger than the MiMo-V2-Flash model Xiaomi released earlier.
Performance and Pricing
MiMo-V2-Pro scores approximately 50 on the Artificial Analysis Intelligence Index, ranking eighth worldwide and second among Chinese large language models, according to community benchmarks compiled during the Hunter Alpha testing period. On ClawEval, a benchmark measuring performance in agent scaffolds, the model scored 61.5, approaching leading frontier models. The model is designed to function as the core intelligence behind AI agents — tools that execute complex tasks with less human input than standard chatbots require.
Xiaomi has priced MiMo-V2-Pro aggressively. For standard prompts up to 256,000 tokens, the model costs one dollar per million input tokens and three dollars per million output tokens. Extending to the full one-million-token context doubles those rates. The company also plans to offer free developer access for one week and to open-source the model once it reaches sufficient stability, according to The Japan Times.
Market Reaction and Strategic Context
Xiaomi’s Hong Kong-listed shares jumped as much as 5.8 percent on March 19, making it the best performer on the Hang Seng Tech Index that day, Bloomberg reported. CEO Lei Jun has stated that Xiaomi will invest more than 16 billion yuan (approximately $2.3 billion) in AI research in 2026, with at least 60 billion yuan ($8.72 billion) committed over the next three years.
The company’s pivot into AI follows a pattern of aggressive diversification. Xiaomi launched its first AI model, MiMo, in April 2025, and its smart EV, AI, and new initiatives segment achieved profitability for the first time in Q3 of that year, generating 29 billion yuan in revenue — a 199 percent year-over-year increase. Xiaomi president Lu Weibing said the company’s AI returns had far exceeded expectations.
The Hunter Alpha episode illustrates a shift in the competitive dynamics of the Chinese AI industry. Where attention has largely focused on dedicated AI labs like DeepSeek and Alibaba’s Qwen team, Xiaomi’s stealth entry demonstrates that hardware conglomerates with deep engineering talent and manufacturing scale are emerging as credible competitors in the foundation model space.