Anthropic, already engaged in a dispute with the government regarding the Pentagon, faced a new directive on June 12th: to restrict foreign access to its recently launched AI models, Fable 5 and Mythos 5. These models, unveiled just days earlier on June 9th, were touted by Anthropic as groundbreaking. The company stated that “Fable 5’s capabilities exceed those of any model we’ve ever made generally available,” and that Claude Mythos 5 shared the same underlying model, albeit “with the safeguards lifted in some areas.”
Reports indicate that this order emerged following discussions between Amazon and the White House. These conversations reportedly centered on findings by researchers who claimed to have discovered methods to prompt Fable 5 into providing information potentially exploitable in cyberattacks.
In response, Anthropic promptly disabled access to both models for all customers. The company issued a statement declaring, “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.”
Over the weekend, Anthropic abruptly took its most advanced AI models offline at the behest of Washington. The American company conveyed that it had little alternative after the White House mandated that it block access for all foreign nationals, a restriction that even extended to its own employees. Internationally, this incident served as a stark reminder that the U.S. not only leads in frontier AI but also exerts significant governmental authority over its usage.
The Trump administration's intervention was characterized by its swiftness, broad scope, and minimal prior warning or explanation. This unprecedented shutdown of the Fable 5 and Mythos 5 models—which had already incorporated safeguards limiting their application in “high-risk areas”—lent new weight to long-standing concerns about over-reliance on the U.S. for critical technologies. It also provided fresh impetus for politicians, governments, and corporations advocating for their own independent leadership in AI development.
A new report from Semafor suggests that the White House’s decision to impose export restrictions on Anthropic’s Mythos was partly fueled by anxieties that a group with ties to China might have accessed the model. Should the Chinese government indeed gain access to Mythos 5 or Fable 5, it would pose a grave national security risk. The report highlighted the potential for reverse engineering the model through distillation, a technique where a “student” AI is trained using a more advanced model to replicate its behavior.
However, the White House has not confirmed this report. Furthermore, a post on X by Trump adviser David Sacks did not reference China, instead focusing on the reported susceptibility of Fable and Mythos to jailbreaking—a claim Anthropic has denied. Anthropic has not responded to requests for comment, though a spokesperson informed Semafor that government officials did not mention China during discussions regarding export controls.
According to The Wall Street Journal, the export control directive that compelled Anthropic to cut off access to Fable 5 and Mythos 5 was partly initiated by cybersecurity research conducted by Amazon and subsequent conversations between Amazon CEO Andy Jassy and the White House. The Journal's report indicates that Amazon’s research claimed to have successfully prompted Fable 5, through a series of commands, to generate information that could be utilized in cyberattacks. Amazon has yet to comment on these findings.
The government’s decision to prohibit foreign nationals from using the models came shortly after Jassy shared Amazon’s findings. This directive created internal complications for Anthropic, as many of its researchers are foreign-born, effectively barring them from accessing their own company's product.
On Friday evening, citing national security concerns, the government officially ordered Anthropic to block access to Fable 5 and Mythos 5 for all foreign nationals, both within and outside the U.S., a mandate that included Anthropic’s own employees. To comply with these demands, the company has completely halted access to the models for all its customers.
In an official statement, Anthropic affirmed its compliance with the order but noted that the government “did not provide specific details of its national security concern.” The company further asserted that any evidence of a potential jailbreak was conveyed verbally, and that the identified vulnerabilities were minor and comparable to those found in other models, including GPT 5.5.
Prior to these events, Anthropic had just announced Claude Fable 5, promoting it as the most powerful AI model it had ever made widely available.
The company had highlighted Fable 5’s “exceptional performance in software engineering, knowledge work, and vision,” noting that its superiority over competing models became more pronounced with the increasing length and complexity of tasks.
The Editorial Staff at AIChief is a team of professional content writers with extensive experience in AI and marketing. Founded in 2025, AIChief has quickly grown into the largest free AI resource hub in the industry.