Anthropic Suspends Claude Fable 5 and Mythos 5 Following US Government Directive

 

Artificial intelligence company Anthropic has suspended access to two of its newest AI models, Claude Fable 5 and Mythos 5, after receiving a government directive tied to national security concerns.

 

The decision comes only days after Anthropic launched Claude Fable 5, a public version of its more advanced Mythos 5 system. The company announced that it had been instructed to restrict access to foreign nationals, a move that ultimately forced the suspension of both models while it works to comply with the order.

 

According to Anthropic, the government did not identify a specific security breach or cyberattack linked to the models. Instead, the company said authorities appeared concerned about a reported technique that could potentially bypass some of the model's built-in safeguards.

 

The issue centers on what is commonly known as "jailbreaking" — a method used to circumvent restrictions placed on artificial intelligence systems. Researchers often use jailbreak techniques to test whether an AI model can be manipulated into generating restricted outputs or performing actions beyond its intended limits.

 

Anthropic said it reviewed demonstrations of the reported technique and found that it was capable of identifying only a small number of previously known vulnerabilities. The company argued that those vulnerabilities were relatively minor and could also be discovered by other publicly available AI systems without bypassing safety protections.

 

 

The suspension marks an abrupt halt for a model that had attracted significant attention across the technology industry.

 

Before its public release, Mythos 5 was made available only to a limited group of organizations participating in security testing programs. Anthropic had repeatedly described the model as one of the most capable AI systems it had ever developed, particularly in cybersecurity-related tasks such as software analysis, vulnerability detection, and code review.

 

The company previously restricted access because of concerns that the model's capabilities could potentially be misused if adequate safeguards were not in place.

 

When Fable 5 was introduced, Anthropic emphasized that the public version contained additional protections designed to prevent malicious use. The company said extensive guardrails had been built into the model to reduce risks associated with cyber exploitation, software attacks, and unauthorized access attempts.

 

Even before launch, however, the model generated debate among policymakers, cybersecurity experts, and technology executives.

 

Supporters argued that advanced AI tools could help organizations strengthen digital defenses by identifying weaknesses faster than traditional methods. Critics warned that increasingly powerful models could eventually lower the technical barriers for cybercriminals seeking to exploit software vulnerabilities.

 

The latest suspension comes at a time when Anthropic is already facing scrutiny from parts of the US government.

 

In recent months, former President Donald Trump publicly criticized the company, while Defense Secretary Pete Hegseth reportedly designated Anthropic as a "supply chain risk." 

 

The designation raised concerns about the use of Anthropic's technology within government-related operations and drew attention because such classifications have historically been associated with foreign companies rather than US-based firms.

 

Anthropic responded by filing a lawsuit against the Pentagon, challenging the designation and arguing that the restrictions were unjustified.

 

A federal judge later ruled that the Pentagon's directive could not be enforced while legal proceedings continue. As a result, government agencies and military contractors remain able to use Anthropic's technology pending the outcome of the case.

 

The suspension of Fable 5 and Mythos 5 adds another layer of uncertainty for Anthropic at a time when competition in the artificial intelligence sector is intensifying. The company has been positioning its Claude family of AI models as a direct competitor to systems developed by OpenAI and Google, while simultaneously promoting stronger safety standards and responsible deployment practices.

 

For now, both Fable 5 and Mythos 5 remain unavailable as Anthropic works to address the government's concerns and determine whether access can eventually be restored.

 

The development highlights the increasing scrutiny facing advanced AI systems as governments around the world seek to balance innovation with concerns surrounding cybersecurity, public safety, and national security.