China's Kimi K3 AI ignites US debate on competition and security

The launch of Kimi K3, a powerful open-source AI model by Chinese startup Moonshot AI, has intensified the technology rivalry between the United States and China, triggering accusations of intellectual property theft and prompting a debate in Washington over how to respond.

What is Kimi K3?

Kimi K3, released on July 16, has impressed with its performance, rivaling top US models like OpenAI's GPT-5.6 and Anthropic's Claude Fable 5 in several benchmarks, especially in programming Source: elespanol. With 2.8 trillion parameters, it is China's largest open-source model, available for free download and modification Source: lavozar. Demand was so high that Moonshot had to pause new subscriptions after 48 hours Source: heraldomx. However, according to consultancy Artificial Analysis, Kimi K3 still lags behind the most advanced US models in overall capability Source: lavozar.

US response and accusations

US officials have reacted sharply. Treasury Secretary Scott Bessent warned of potential sanctions if evidence emerges that Chinese firms used "distillation"—training a model on another's outputs—to copy US technology Source: lavozar. Michael Kratsios, White House science and technology director, explicitly accused Moonshot of copying Anthropic's Fable model Source: excelsiormx. The Commerce Department is investigating whether Chinese companies evaded export controls on Nvidia chips Source: excelsiormx.