On June 12, 2026, the US government ordered Anthropic to disable its newest AI models—Claude Fable 5 and Mythos 5—worldwide, citing export controls. The next day, Beijing time, Chinese AI firm Zhipu released open-source GLM-5.2 under an MIT license, offering frontier intelligence freely to global developers.
According to Anthropic's official statement, the directive arrived at 5:21pm ET, barring access by any foreign national, including Anthropic's own employees. The company disabled the models globally, citing an inability to verify nationality, and is working to restore access. It noted that the demonstrated vulnerabilities are present in other publicly available models and disagreed with the directive.
On June 13, 2026, Beijing time, Zhipu released GLM-5.2, as announced on its official WeChat account. The model opens to all GLM Coding Plan tiers, with an API launch next week. Founder Jie Tang posted on X: "Frontier intelligence belongs to everyone." The contrast indicts US containment: Washington locks its best AI behind walls, while Beijing's open-source release accelerates Chinese self-reliance and positions China as the champion of open innovation.