Anthropic has been forced to suspend all access to its Fable 5 and Mythos 5 AI models following a government-issued export control directive citing national security concerns, a move the company says it is complying with under protest.
The directive, received on June 13, 2026, at 5:21 PM ET, prohibits any foreign national, whether located inside or outside the United States, from accessing the two models, including foreign national Anthropic employees.
To ensure full compliance, Anthropic has disabled Fable 5 and Mythos 5 globally for all customers. Access to all other Anthropic models remains unaffected.
Anthropic Blocks Fable 5 and Mythos 5 Access
The government’s directive did not provide specific details of the national security risk. Still, according to Anthropic, it appears to stem from a reported method for bypassing or “jailbreaking” Fable 5’s safety guardrails.
A demonstration of the technique was reviewed by Anthropic, which found it was used to identify a small number of previously known, minor vulnerabilities.
Critically, Anthropic says those vulnerabilities are already discoverable using other publicly available models, suggesting no unique uplift from the Fable 5 bypass.
“To date, the government has only given us verbal evidence of a potential narrow, non-universal jailbreak, which essentially consists of asking the model to read a specific codebase and fix any software flaws,” Anthropic stated in its public disclosure.
The company further noted it reviewed what it believes to be the underlying report driving the directive and validated that the level of capability demonstrated is “widely available from other models,” explicitly referencing OpenAI’s GPT-5.5 cybersecurity evaluation disclosures.
Before launch, Anthropic worked with the US government, the UK AI Safety Institute (AISI), multiple private third-party organizations, and internal red teams to stress-test Fable 5’s safeguards for thousands of hours.
Those evaluations confirmed that Fable’s guardrails are substantially more effective than any previously deployed frontier model, and no tester has yet discovered a universal jailbreak capable of broadly bypassing the model’s safeguards across a wide range of cyber tasks.
Acknowledging that perfect jailbreak resistance is not achievable for any model provider today, Anthropic adopted a defense-in-depth strategy for Fable 5.
The approach is designed to keep non-universal jailbreaks narrow in scope, make universal jailbreaks prohibitively expensive to develop, and combine technical safeguards with continuous monitoring to detect and shut down successful attacks rapidly.
Anthropic also enforces a 30-day customer data retention policy for Mythos-class models, a measure the company says carries real costs with customers but is essential for ongoing jailbreak research and mitigation.
While complying with the legal directive, Anthropic publicly disputed the government’s reasoning.
The company stated it has not received disclosure of a jailbreak that led to a harmful outcome, characterizing the disclosed findings as either entirely benign responses or minor findings that provide no Mythos-specific uplift over capabilities already available in other deployed models.
Anthropic warned that applying this recall standard broadly across the industry would essentially halt all new model deployments for every frontier AI provider.
The company reiterated its position that governments should retain authority to block unsafe AI deployments. Still, it argued such actions must follow a process that is transparent, fair, clear, and grounded in technical facts, criteria it says this directive fails to meet.
Anthropic apologized to affected customers for the disruption and stated it is working to restore access as quickly as possible, with a commitment to share additional technical details within 24 hours.
Follow us on Google News , LinkedIn and X to Get More Instant Updates. Set Cyberpress as a Preferred Source in Google.