A new artificial intelligence tool introduced by the Chinese firm Moonshot AI, which is based in Beijing, appears to have once again eliminated the lead in this technology claimed by the likes of Anthropic and Open AI in the United States. The new model just introduced by Moonshot AI is called Kimi K3; the AI evaluator called Arena says that Kimi K3 beat the best models of the two U.S. firms in what are called “front-end coding tests.” Moreover, as last year with the Chinese model Deepseek, Kimi K3 is far cheaper to buy than any American large-language model. And it is not being kept behind a proprietary wall, but rather made available, as of its official release July 27, to other companies and other governments, to test or improve their own models. This is reported July 17 by Axios, and at greater length in the tech publication Futurism.
As to why this has “jolted the foundations of America’s AI boom,” as Axios put it: The overwhelming assumption of primarily war applications, so that the leading position in AI means future military domination of one nation over all others, underlies this competition. Thus, as important as the competitive surprises by China, is the major effort being led by the Vatican of Pope Leo XIV, to define the uses of AI as “disarming” the arms race.
