US lifts restrictions on Anthropic's advanced AI models

Lead

The United States has lifted export controls on Anthropic's most advanced AI models, Fable 5 and Mythos 5, allowing the company to restore access globally, Anthropic announced on Tuesday. The decision comes less than three weeks after the US government ordered a suspension of the models over national security concerns.

Coverage Comparison

The move was reported by multiple outlets, with variations in focus and detail. The Jerusalem Post, citing a source from Anthropic, reported that Fable 5, the company's most powerful AI model, would be redeployed outside the US soon after more than two weeks of being banned. It highlighted that Anthropic disagreed with the ban, citing "verbal evidence of a potential narrow, non-universal jailbreak." The outlet also noted that the Trump administration had banned access to Fable 5 on June 12 due to concerns about its potential to identify software vulnerabilities.

Dawn emphasized that Anthropic agreed to cooperate with the US government on protocols for Mythos, Fable, and future models, and that the company would proactively detect and address security risks. The report also noted that OpenAI limited the rollout of its GPT-5.6 model to vetted partners at the US government's request.

Deutsche Welle focused on the cooperation between Anthropic and the US government, and included a comparison by CIA Director John Ratcliffe of advanced AI models to nuclear weapons. The outlet also reported that European lawmakers described the US government's ability to shut off access as a "kill switch."

France 24 – English reported that the US government lifted the restriction on Fable 5 and Mythos 5, and that Anthropic will restore access globally. It also mentioned the CIA director's nuclear weapons comparison and OpenAI's limited rollout of GPT-5.6.

The Jerusalem Post, which provided the most detailed account of the ban, reported that the ban was lifted after Anthropic addressed security concerns and that the company will redeploy Fable 5 outside the US soon. It also noted the Trump administration's concerns about Fable 5's potential to identify software vulnerabilities.

Key Claims

Anthropic announced on Tuesday that it had received notice from the Department of Commerce that export controls on Claude Fable 5 and Mythos 5 had been lifted, and that it would begin restoring access the following day. This was reported by multiple outlets, including Dawn, Deutsche Welle, France 24, and the Jerusalem Post.

According to the Jerusalem Post, the ban was imposed on June 12 after the Trump administration expressed concerns that Fable 5 could be used to identify software vulnerabilities. Anthropic disagreed with the decision, citing the alleged jailbreak evidence. The company argued that a "narrow potential jailbreak" should not justify recalling a model deployed to hundreds of millions of people.

Anthropic implemented a new safeguard to block a technique it believes the government viewed as a method of bypassing Fable 5's security, as reported by the Jerusalem Post. Commerce Secretary Howard Lutnick, in a letter seen by the Post and reported by Dawn, confirmed that restrictions were lifted after Anthropic agreed to proactively address security risks and cooperate on protocols for the models and future ones. Lutnick also said the department could reimpose restrictions if circumstances change or if Anthropic fails to meet its commitments.

CIA Director John Ratcliffe, as quoted by both Deutsche Welle and France 24, compared advanced AI models to nuclear weapons, reflecting the security concerns that have driven the US government's oversight of AI.

OpenAI, according to Dawn and France 24, limited the rollout of its GPT-5.6 model to a set of vetted partners at the request of the US government, a parallel to Anthropic's situation.

Perspectives

• Anthropic: The company welcomed the lifting of restrictions, stating it would restore access to its models. It had previously disagreed with the ban, calling it based on "verbal evidence of a potential narrow, non-universal jailbreak" and arguing that a narrow potential should not lead to a recall of a widely used model.

• US Government: The Commerce Department and the Trump administration, citing national security concerns, initially imposed the restrictions to prevent misuse of advanced AI models. After negotiations, the restrictions were lifted following Anthropic's agreement to address risks and cooperate on security protocols. Officials reserved the right to reimpose restrictions if needed.