Anthropic Suspends Access to AI Models Fable 5 and Mythos 5 Following US Government Directive

Anthropic, a leading AI technology company, has taken a significant step by suspending access to its advanced AI models, Fable 5 and Mythos 5, for users worldwide. This action comes in response to a directive from the US government, instructing the company to block access for any foreign national.

US Government Directs Anthropic to Restrict Access to Fable 5

The directive, received by Anthropic at 5:21pm ET on June 12, cites “national security” concerns as the basis for barring access to both models for foreign nationals, regardless of their location. This includes Anthropic’s own employees who are foreign nationals.

As a result of this directive, Anthropic has been compelled to disable both Fable 5 and Mythos 5 for all customers in order to comply. It’s important to note that other models offered by Anthropic, such as Claude Opus 4.8, remain unaffected by this directive.

This development comes at an awkward time as Anthropic recently launched Fable 5 on June 9, offering it for free to Pro, Max, and Enterprise customers until June 22. The sudden suspension of a model that was made available to millions just three days ago has impacted all users.

Fable 5 serves as the secure version of Mythos 5, with both models sharing the same core structure. However, Fable 5 includes additional security measures to block or redirect sensitive queries related to cybersecurity, biology, and chemistry. On the other hand, Mythos 5 is limited to authorized government cyberdefenders and partners in life sciences.

Anthropic has communicated to developers that new sessions will default to the user’s primary model or Opus 4.8, existing Fable 5 sessions will encounter errors, and requests to the Fable 5 model via the platform will be unsuccessful. Integration partners have been advised to transition to alternative models.

Kanishka Narayan MP, UK’s Minister for AI and Online Safety, has highlighted the impact of this directive on customers in both the US and UK. The Minister has emphasized the importance of technological sovereignty, referencing the government’s substantial investment in AI chip technology.

Anthropic has expressed its belief that the directive stems from concerns regarding a potential method to bypass security measures in Fable 5. The company conducted an assessment of a demonstration and found only minor, known issues, which are also present in publicly available models without any circumvention.

The company stated, “Our understanding is that one potential jailbreak was shared with the government.” Anthropic is complying with the directive while maintaining that the identified security vulnerability does not warrant the complete suspension of a widely deployed commercial model.

Anthropic has pointed out that similar capabilities are readily accessible elsewhere, citing OpenAI’s GPT-5.5 as an example, which is utilized by security practitioners on a daily basis.

While acknowledging the directive, Anthropic is actively working to restore access to Fable 5 and Mythos 5 and has pledged to provide further updates within the next 24 hours.

Security teams identify only 14% of attacks, leaving the majority undetected. Discover how breach and attack simulation can enhance your security posture in the Picus whitepaper.

See also  Palo Alto GlobalProtect VPN Vulnerability Exploited in Recent Cyber Attacks

Download the whitepaper now for valuable insights.