Anthropic to Meet with U.S. Government Officials Today to Negotiate Model Export Control Dispute
N.R. Finch
Anthropic senior staff will meet Trump administration officials in Washington on Monday, seeking to reverse a Friday export-control order that invoked national-security powers to bar its newest models Fable 5 and Mythos 5 from all foreign nationals — forcing the company to cut off every customer.
What exactly did the government order?
On Friday, Anthropic received an export-control directive citing "national-security powers," demanding the company suspend access to Fable 5 and Mythos 5 for "any foreign national, whether inside or outside the United States."
To comply, Anthropic shut down both models for all customers — not just foreign users, but U.S. users too.
This means → the order goes far beyond a typical export restriction. It effectively freezes a live commercial product already serving a global user base.
Why does Anthropic say the order is unjustified?
According to a person familiar with the matter, Anthropic worked with government agencies to test the models before launch and received deployment approval. No national-security concerns were communicated before Friday's directive.
The company believes the government's worry centers on a "potential, narrow, non-universal jailbreak" — a scenario where users bypass cybersecurity guardrails and ask Fable 5 to read a specific codebase and fix software flaws.
Anthropic stated directly: "We disagree with using the discovery of a narrow potential jailbreak as grounds for recalling a commercial model deployed to hundreds of millions of people."
In plain terms = Anthropic's position is that a narrow vulnerability should not trigger a full product recall for a model already live at massive scale.
What are Fable 5 and Mythos 5 built on?
Both models are built on Anthropic's Claude Mythos Preview — a model designed to find software security vulnerabilities, previously restricted to a select group of companies under a cybersecurity initiative called Project Glasswing.
Anthropic says Fable 5 and Mythos 5 lead on multiple industry benchmarks. Mythos 5 remains limited-access; Fable 5 was opened to enterprise clients and paid subscribers.
The company's rationale for the broader Fable 5 release: new safety measures can block responses in specific high-risk domains, including cybersecurity and biology.
What does this reveal about a deeper rift?
This is not an isolated clash. In March, the U.S. Department of Defense designated Anthropic a supply-chain risk, barring defense contractors from using the company's technology on national-security grounds.
Defense Secretary Pete Hegseth responded to the latest directive on X on Saturday, saying "every day that passes" confirms that blacklisting Anthropic was "the right call."
This reflects a relationship that has shifted from partnership to adversarial — a company built on safety research now treated by the government as a security threat.
What happens if this standard applies industry-wide?
Anthropic warned that if a "narrow potential jailbreak" can trigger a product recall, applying that standard across the industry would "essentially prevent all frontier model providers from deploying all new models."
This means → the dispute is no longer about one company or one product. It is about where the government sets the safety-review threshold for AI models — a precedent that would bind OpenAI, Google, and every other frontier lab.
Content is for reference only, not financial advice.