Moonshot AI denied online claims that its founder and employees had been taken away, calling the information fabricated and malicious and saying it had reported the matter to police and would pursue legal action. The denial follows a wider dispute over Kimi K3, which Moonshot released on July 16 as the world’s largest free AI model. The model quickly strained capacity, prompting the company to stop accepting new subscriptions four days later while paying users retained access. On Sept. 11, Moonshot deployed the smaller K2.8 Preview across its coding platform, and some K3 requests appeared to be answered by that model instead. Anthropic then accused Moonshot of routing nearly 300,000 Kimi requests through 5,380 fake accounts to Claude during a 10-day period and presenting Claude’s answers as Kimi’s work. Anthropic said one user was probably affiliated with the military and had uploaded footage from hundreds of security cameras in Chengdu. The allegations rely on Anthropic’s own logs and have not been tested by a regulator; no newspaper or agency has confirmed an arrest or investigation.