Key development
Anthropic announced on Friday that it will "abruptly disable" its top-tier AI models after receiving a U.S. government export control directive that ordered the company to cut off access for all foreign nationals to its newest models, Fable 5 and Mythos 5. The company said the directive was issued without a detailed written explanation of the national security concern motivating the move.
Government concern and company response
According to Anthropic, the Commerce Department's action reflects a belief that there exists a way to bypass safeguards - commonly described as a "jailbreak" - which could allow Fable 5 to be used to identify software vulnerabilities. Anthropic said the government provided only verbal evidence of what it characterized as a "potential narrow, non-universal jailbreak." The firm pushed back against the breadth of the government response, arguing that a limited, theoretical bypass should not trigger recalling a commercial model deployed to hundreds of millions of users.
Anthropic framed the order as an overbroad application of a narrowly scoped finding, warning that if the same standard were applied across the industry, it would effectively halt new model deployments for all frontier model providers. The company said it believed there had been a "misunderstanding" and that it is working to restore access to the affected models as soon as possible.
Operational impact
As a result of the directive, Anthropic said it must immediately disable Fable 5 and Mythos 5 for all customers to ensure compliance. It specified that access to all other Anthropic models will not be affected. Amazon Web Services said Anthropic had asked it to revoke access to the models for "all users in all regions." A U.S. official confirmed the Commerce Department had issued an export control directive to suspend all access to Fable 5 and Mythos 5 by foreign nationals.
Security and industry context cited by both sides
Anthropic has maintained that it worked with the U.S. government and other parties on safety ahead of the Fable rollout. The company also noted that rival providers' models exhibit similar capabilities in surfacing minor bugs in code. The firm said Fable 5 was accompanied by guardrails barring its use in higher-risk areas like cybersecurity, and acknowledged that some users had complained those protections were "overly broad."
Government officials have signaled heightened concern that advanced "Mythos-class" models could accelerate the development of sophisticated cyberattacks if misused, particularly in industries such as banking where complex, interdependent legacy systems are common. The Commerce Department's directive reflects a broader U.S. effort to constrain potential adversaries' access to advanced AI capabilities by restricting use of the models themselves, rather than focusing solely on the hardware and tools that power AI.
Political and regulatory backdrop
The directive represents a further escalation in tensions between Anthropic and parts of the U.S. government. Earlier this year, tensions worsened after Anthropic declined requests from the U.S. military to permit its AI models to be used for domestic surveillance or fully autonomous weapons systems. In response, the government placed Anthropic on a supply chain blacklist that was scheduled to take effect later in the year. The new export control directive arrives amid signs that some disagreements between administration officials and the company had shown signs of easing in parts of the government prior to this action.
Anthropic also disclosed that it confidentially filed for a U.S. initial public offering last month. The company said the recent government action did not adhere to principles of fair and fact-based regulation, even though it had previously advocated for stronger U.S. oversight of AI and the authority to block models that posed unacceptable risks.
Reactions from government and others
The Pentagon’s chief information officer, Kirsten Davies, posted in support of the decision, saying the Defense Department prioritized national security and that some considerations transcend commercial interests and company valuations. A former White House official involved in the AI Action Plan commented that the order implies non-U.S. users would be restricted from using Anthropic’s latest models and suggested users may need to prove citizenship to access them.
People and personnel considerations
Several notable Anthropic employees, including co-founder Chris Olah, AI researcher Andrej Karpathy and philosopher Amanda Askell, were born outside the United States. Anthropic declined to comment on whether such staff would lose access to the affected AI models under the export control directive.
Broader implications and company stance
Anthropic characterized the government's evidence as verbal and narrow, and argued that the directive's effects are disproportionate to the claimed risk. The company warned that applying the same measure industry-wide could stall frontier model deployment. At the same time, officials cited national security priorities in supporting the directive, and highlighted concerns about misuse of advanced models to speed up cyberattacks against critical sectors.
This report presents the information Anthropic and government officials have made public about the directive and the company's response. Anthropic said only Fable 5 and Mythos 5 are affected and that other models remain accessible.