A breakdown of the June 2026 export control directive, the national security concerns behind it, and Anthropic's response
Quick Summary
On the evening of Friday, June 12, 2026, Anthropic disabled global access to two of its newest AI models Fable 5 and Mythos 5 after receiving a directive from the US Commerce Department invoking export control authorities. The order instructed Anthropic to prevent any foreign national from accessing the company's latest and most powerful models, and the company said it had to disable the models for all customers not just foreign nationals to comply. Access to other Claude models, including Claude Opus 4.8, was not affected.
Timeline of Events
June 10, 2026
Anthropic CEO Dario Amodei published a policy essay describing Mythos as a clear example of how quickly frontier AI capability and risk were advancing, and called for mandatory third-party testing of models across four risk categories: cybersecurity, biological weapons, loss of control, and AI-accelerated R&D. This came just two days before Anthropic announced it had received the government's directive.
June 12, 2026, ~5:21 PM ET
Anthropic received a letter from Commerce Secretary Howard Lutnick, addressed to CEO Dario Amodei, stating that Mythos 5 and Fable 5 would be subject to export controls covering any location outside the US and all foreign persons within the country.
Same evening
Anthropic announced it had to abruptly disable both models for all customers worldwide to ensure compliance, since the directive applied to foreign nationals "whether inside or outside the United States," including the company's own non-citizen employees.
What Are Fable 5 and Mythos 5?
The two models are closely related but serve different audiences:
•Fable 5 A newly released, publicly available model. It is built on Mythos technology, but its cybersecurity and biotechnology-related capabilities are restricted. It was the first time Anthropic had released such an advanced model to the public, made possible by new safeguards designed to block responses in specific high-risk areas.
•Mythos 5 A non-public, full-capability version. It was intended for use only by government agencies and select corporate partners to help harden their own systems. This model was tied to an earlier cybersecurity initiative called Project Glasswing, and Anthropic said it never planned to make it generally available, limiting its rollout to a select group of organizations.
The Government's Stated Concern
The Commerce Department's directive was framed around national security, but details were thin at first:
•The letter did not specify what the national security concern actually was.
•An administration official told Axios that the decision followed a claim from another company that it had managed to jailbreak Mythos, which alarmed officials about potential national security risks.
•The administration reportedly tried to get Anthropic to pause the release of the new models before resorting to the export control letter.
•According to Anthropic's own review, the technique in question amounted to a limited capability allowing the AI to review a specific codebase and correct errors in it — not a sweeping safety bypass.
•Under Commerce's letter, a license is now required for any export, re-export, or domestic transfer of these models.
Anthropic's Response
Anthropic pushed back firmly while complying with the order:
•The company said the reported vulnerabilities appeared to be relatively simple and were discoverable by other publicly available AI models as well.
•It argued the jailbreak was narrow — capable of unlocking Mythos's cybersecurity functions in one specific instance rather than defeating all of Fable 5's safeguards universally — and that the same technique could likely extract similar behavior from competitor systems, including OpenAI's GPT-5.5, which face no comparable export restrictions.
•In a public statement, Anthropic wrote: "We disagree that the finding of a narrow potential jailbreak should be cause for recalling a commercial model deployed to hundreds of millions of people."
•On social media, the company apologized for the disruption and said: "We believe this is a misunderstanding and are working to restore access as soon as possible."
•Anthropic also said the government had so far provided only verbal evidence regarding the issue.
Why This Is a Big Deal
Several factors make this episode unusual:
1. A first for the industry. This appears to be the first time a leading AI company has taken a publicly deployed model offline because of intervention from the federal government.
2. A first for export controls. To the best of available knowledge, this marks the first time the United States has issued an export control directive specifically targeting LLM access.
3. Global reach. The directive doesn't just affect the US — it extends to close allies including the UK, Canada, Australia, and New Zealand, restricting access for their nationals regardless of where they reside.
4. Policy backdrop.The move follows an executive order earlier in the month from the Trump administration requiring the most advanced AI models to undergo testing before deployment, and Anthropic already had a pre-deployment testing partnership with Commerce's Center for AI Standards and Innovation.
5. Timing for Anthropic. The shutdown came shortly after Anthropic confidentially filed for a public stock market listing, and reporting noted the news triggered a sharp decline in pre-IPO market valuations.
What Happens Next
According to Axios's source, the restriction isn't necessarily permanent: an administration official indicated the models would need to stay locked down until the government's national security apparatus is sufficiently hardened, a process that could take a few weeks. As of the directive, both companies and outside observers were still waiting for the government to provide fuller documentation of the specific vulnerability that triggered the action.
Sources
Anthropic, "Statement on the US government directive to suspend access to Fable 5 and Mythos 5"
