Why US Blocked Access to Anthropic Fable 5 and Anthropic Mythos 5 Models
The US government has ordered an immediate suspension of Anthropic’s Fable 5 and Mythos 5 models, citing national security concerns over a potential jailbreak. Anthropic is complying with the directive while disputing the necessity of the recall, arguing that the identified vulnerabilities are minor and standard across the current landscape of AI technology.
The United States government has issued an urgent export control directive mandating the immediate suspension of Anthropic’s Fable 5 and Mythos 5 artificial intelligence models. This regulatory intervention, enacted on June 12, 2026, requires that all access to these specific models be blocked for foreign nationals, including those employed by Anthropic, both within and outside the United States.
Anthropic confirmed it received the directive at 5:21pm ET, noting that the order cites national security authorities. While the company is complying with the mandate to disable Fable 5 and Mythos 5 for all customers, it maintains that the decision appears to be based on a misunderstanding regarding the model's safety and current security capabilities. US Government Orders Anthropic To Suspend Access to Claude Fable 5, Mythos 5 Over Export Control Directive.
Anthropic Statement Over US Government Blocking Fable 5 and Mythos 5
The US government, citing national security authorities, has issued an export control directive to suspend all access to Fable 5 and Mythos 5 by any foreign national, whether inside or outside the United States, including foreign national Anthropic employees.
The net effect of…
— Anthropic (@AnthropicAI) June 13, 2026
US Government Concerns and Security Allegations Over Anthropic models
The directive stems from government concerns regarding a potential "narrow jailbreak" method discovered within Fable 5. According to Anthropic, the government suggests this method allows users to bypass safety guardrails to identify software vulnerabilities. However, the company argues that these findings are minor and comparable to capabilities already present in other publicly available models, such as GPT-5.5.
Anthropic stated that it has reviewed the specific technique mentioned by the authorities. The company maintains that the level of capability displayed in the demonstration is widely used by cybersecurity defenders to maintain system integrity. The organisation expressed concern that the standard applied by the government could, if consistently implemented, halt future deployments for all frontier model providers.
Anthropic Highlights 'Defense in Depth' Strategy
In response to the recall, Anthropic highlighted its "defense in depth" strategy, which was central to the launch of Fable 5. This approach combines sophisticated safety classifiers with thorough monitoring systems designed to detect and mitigate misuse. The company noted that it has required a 30-day data retention policy for Fable users specifically to facilitate research into jailbreak attempts.
The company emphasised that it conducted thousands of hours of red-teaming prior to the release of Fable 5 in collaboration with the UK Artificial Intelligence Safety Institute and other private organisations. Anthropic remains committed to its safety posture, noting that no universal jailbreak—which would allow broad bypasses of system capabilities—has yet been identified by testers.
The sudden removal of these models affects a significant number of developers and enterprises that had begun integrating Fable 5 into their workflows since its June 9, 2026, launch. The model was designed to handle complex, long-term projects and featured a 1-million-token context window. Anthropic has apologised to its user base for the disruption caused by the abrupt compliance measure. Anthropic Valuation Surpasses OpenAI Hitting USD 965 Billion, AI Company Secures USD 65 Billion in Series H Funding.
Looking forward, Anthropic has stated it is working to resolve the matter with the government as quickly as possible. The company advocates for a regulatory process that is transparent, fair, and grounded in technical facts, suggesting that current actions against its models do not align with those principles. Further details are expected within the next 24 hours.
(The above story first appeared on LatestLY on Jun 13, 2026 09:27 AM IST. For more news and updates on politics, world, sports, entertainment and lifestyle, log on to our website latestly.com).
