US accuses China's Moonshot AI of using Anthropic's Fable to build K3 model

China's Moonshot AI has been accused of using distillation techniques on Anthropic's most advanced large language model, Fable, to develop its latest K3 model. The allegation was made by White House Office of Science and Technology Policy Director Michael Kratsios on Wednesday.

In a post on X, Kratsios said, "We have information that Moonshot AI distilled Anthropic's Fable for the development of its K3 model."

The allegation comes months after the US State Department, according to Reuters, ordered a global diplomatic push in April to highlight what it describes as widespread efforts by Chinese companies—including Moonshot AI—to steal intellectual property from US artificial intelligence labs.

What is Distillation

Distillation is a technique in which smaller AI models are trained using the outputs of larger, more capable models. The process is widely used to reduce the cost and computing resources required to build powerful AI systems. Moonshot AI recently unveiled Kimi K3, a 2.8-trillion-parameter model that the company claims is the world's largest open-weight AI model and delivers performance approaching Anthropic's frontier Fable model.

More from Tech

The allegations also stand in contrast to recent remarks by Nvidia CEO Jensen Huang, who argued that Chinese AI models should not be viewed as a threat. Speaking to Axios, Huang said there was "zero possibility" that Chinese AI companies would overtake or drive US firms out of the market. "There's no scenario where China runs U.S. companies off the road. Zero possibility," Huang said, while advocating for broader adoption of open AI models, including those developed in China.