Xiaomi Confirms Hunter Alpha as Its AI Model, Quashing DeepSeek V4 Rumors
A powerful artificial intelligence model that emerged anonymously on the developer platform OpenRouter last week has been officially identified as belonging to Chinese technology giant Xiaomi. This revelation puts to rest widespread speculation that the model was a stealth test of DeepSeek's highly anticipated next-generation V4 system.
The Mysterious Emergence of Hunter Alpha
The model, named Hunter Alpha, first appeared on OpenRouter on March 11 without any developer attribution, quickly being labeled a stealth model by the platform. Its sudden appearance fueled intense interest in the AI community, particularly due to its advanced specifications and free access, which led many to suspect it was DeepSeek's unreleased V4 model.
DeepSeek's previous releases, such as the low-cost DeepSeek-V3 and R1 models, had triggered a global tech stock selloff last year by challenging the high spending norms of US AI firms. Consequently, anticipation for DeepSeek-V4 has been building, with Chinese media reporting a potential launch as early as April.
Xiaomi's AI Team Steps Forward
On Wednesday, Xiaomi's AI model team, MiMo, led by former DeepSeek researcher Luo Fuli, announced that Hunter Alpha is actually an early internal test build of MiMo-V2-Pro. This flagship model is designed to serve as the brain of AI agents, enabling users to execute complex tasks with minimal human prompts and supervision compared to traditional chatbots.
Luo described the launch as a quiet ambush, noting in an X post on Thursday that the shift from chat to agent paradigms has occurred so rapidly that even her team was taken aback. People ask why we move so fast, she added. I saw it firsthand building DeepSeek R1.
The timing aligns with the rapid adoption of OpenClaw, an open-source agent framework in China. MiMo-V2-Pro will partner with five major agent frameworks, including OpenClaw, to offer a week of free access to developers worldwide. Following the announcement, Xiaomi's Hong Kong-listed shares surged as much as 5.8 per cent on Thursday.
Technical Specifications and Testing
During tests conducted by Reuters last week, the Hunter Alpha chatbot described itself as a Chinese AI model primarily trained in Chinese with a knowledge cutoff date of May 2025—matching DeepSeek's own chatbot. However, it declined to identify its creator, stating only, I only know my name, my parameter scale and my context window length.
Hunter Alpha's profile lists it as a 1-trillion-parameter model, trained using approximately one trillion adjustable values that dictate language processing and response generation. It also boasts a context window of up to one million tokens, allowing it to process or remember extensive text in a single interaction.
Nabil Haouam, an engineer specializing in AI agent systems, highlighted the model's appeal: The combination that stood out was Hunter Alpha's 1-million-token context paired with reasoning capability and free access. Most frontier models with that context window come with real cost at scale.
These features closely matched media expectations for DeepSeek-V4, leading to the initial speculation. Umur Ozkul, who runs independent AI benchmark tests, noted that the timing and capabilities made the connection to DeepSeek understandable.
Stealth Testing Practices in AI
Stealth model launches are common in the AI industry, as platforms like OpenRouter allow developers to test new systems anonymously through a single interface. For instance, an anonymous model called Pony Alpha appeared on OpenRouter in February before Chinese firm Zhipu AI confirmed it as part of its GLM-5 system five days later.
A notice on Hunter Alpha's profile indicated that all prompts and completions are logged by the provider and may be used to improve the model, underscoring the practice of using stealth launches for unbiased feedback. According to MiMo, the model was rapidly adopted after its appearance, surpassing one trillion tokens in total usage and topping OpenRouter's leaderboard charts.
This development marks a significant move by Xiaomi into the competitive AI landscape, leveraging advanced agent capabilities to potentially reshape how AI tools are deployed and utilized globally.
