Anthropic blocks access to the latest AI models by order of the USA

Anthropic blocks access to the latest AI models by order of the USA

By Pieter Haeck
Publication Date: 2026-06-13 10:11:00

The company claimed that the U.S. government imposed export controls because it feared that some of the models’ built-in security measures could be circumvented, a process known as “jailbreaking.”

“So far, the government has only given us verbal evidence of a potentially limited, non-universal jailbreak,” the company said. It added that it does not believe “that the finding of low potential jailbreak should be grounds for recall of a commercial model.”

In early April, Anthropic announced that it had granted access to Mythos Preview to a limited number of trusted technology and cyber companies through the Glasswing project to help them strengthen cyber defenses.

The decision left many non-US governments and entities, such as the EU institutions, fighting for access.

Fable 5, which launched just this week, is described as a “Myth-class model” that the company has added safeguards to make it safe for general use.

“Without safeguards, Fable 5’s capabilities in areas such as cybersecurity could be compromised…”