US Government Orders Anthropic to Suspend Fable 5 and Mythos 5 Over Jailbreak Concerns

The US government moved swiftly on national security grounds, issuing an export control directive that forces Anthropic to cut off all access to its frontier models Fable 5 and Mythos 5 for any foreign national, including the company's own foreign-born employees, whether they're in the country or abroad. The order arrived at 5:21 PM ET and took effect immediately, leaving Anthropic little choice but to disable both models for all customers to stay within the law. Other Anthropic models remain unaffected.

Government's National Security Concern

The directive, delivered by letter, offered no specifics on the underlying threat. In its own account, Anthropic says the government believes it has found a method to bypass, or “jailbreak,” Fable 5. The company reviewed a demonstration of the technique and concluded it only exposed a handful of previously known, minor vulnerabilities. Anthropic argues these flaws are relatively simple and that other publicly available models can find them without any jailbreak at all.

Anthropic's Defense-in-Depth Approach

In its launch blog post, Anthropic laid out a multi-layered security strategy for Fable 5. The company said it built strong safeguards that dramatically lower the odds of misuse, especially for cybersecurity-related tasks, so much so that many users have complained they're too restrictive. Before launch, Anthropic worked with the US government, the UK AI Safety Institute, several outside groups, and its own internal teams on thousands of hours of red-teaming.

The tests indicated Fable's safeguards are substantially more effective than any previous model's. No tester has yet found a universal jailbreak, a technique that would broadly bypass the model's protections and unlock a wide range of cyber capabilities. Still, Anthropic acknowledged that perfect jailbreak resistance is out of reach for any provider today, and that all industry safeguards remain vulnerable to non-universal jailbreaks that can, under specific conditions, extract some cyber information.

“Given that perfect jailbreak resistance does not appear to be possible today, Anthropic adopted a defense in depth strategy with Fable 5,” the company stated. “We aimed to make jailbreaks either narrow (in the case of non-universal jailbreaks) or very expensive to produce (in the case of universal jailbreaks), and to combine this with thorough monitoring to quickly detect and shut down any successful attacks.” That strategy also included a 30-day retention policy for customer data from Fable, a change that carries real costs, the company says, but enables research into jailbreaks and faster mitigation.

Dispute Over the Basis of the Directive

Anthropic disagrees sharply with the government's move. The company says that so far, officials have only offered verbal evidence of a potential narrow, non-universal jailbreak, essentially asking the model to read a specific codebase and fix software bugs. Anthropic reviewed a report it believes underpins the directive and found that the level of capability shown is widely available from other models, including OpenAI's GPT-5.5, and is used every day by defenders keeping systems safe.

“We are complying with the government’s legal directive and are removing access to Fable 5 and Mythos 5 for all users. However, 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,” Anthropic said. “If this standard was applied across the industry, we believe it would essentially halt all new model deployments for all frontier model providers.”

Anthropic reiterated its position that the government should have the power to block unsafe deployments, but only as part of a statutory process that is transparent, fair, clear, and grounded in technical facts. “This action does not adhere to those principles,” the company said.

Impact and Next Steps

Anthropic apologized to customers for the disruption and said it hopes the situation turns out to be a misunderstanding. The company is working to restore access as quickly as it can. The directive marks another flashpoint in the growing tension between national security imperatives and the breakneck pace of frontier AI deployment, and raises fresh questions about what criteria the government should use when it decides to step in.