US Lifts Export Ban on Anthropic's Fable 5 and Mythos 5 Models

The Trump administration has lifted export restrictions imposed in mid-June on Anthropic's two most advanced models, Claude Fable 5 and Mythos 5. Commerce Secretary Howard Lutnick announced the decision on June 30, 2026, and the company began restoring access to customers worldwide starting July 1.

The sudden ban was driven by national security concerns. The US government pointed to a jailbreaking method, a way of bypassing Fable 5's safeguards, that reportedly allowed access to functions restricted because of cyber, biological, and chemical risks. Anthropic downplayed the scale of the problem, saying it involved "a small number of previously known, minor vulnerabilities" rather than a gateway to sophisticated attacks.

Two weeks of negotiations

Between mid-June and the end of the month, Anthropic representatives met with Trump administration officials to work out the terms for restoring access. The company argued that both models, Mythos and Fable, have built-in restrictions on queries related to cybersecurity, biology, and chemistry, and that their main use is meant to be identifying security vulnerabilities, not exploiting them.

The Commerce Department ultimately agreed to lift the ban in exchange for specific commitments from Anthropic. The company is to proactively detect and remove security threats tied to its models, work closely with the government to establish protocols and standards for Mythos, Fable, and future models, and inform the administration of any malicious activity it detects.

Conditional trust

Lutnick's letter included a clear caveat: the restrictions could return if "circumstances change" or if Anthropic fails to honor its commitments. It's a sign that Washington views its relationship with AI companies as a conditional partnership, not a permanent release from oversight.

The priority is deploying the best technology as quickly and safely as possible - Susie Wiles, White House Chief of Staff

Wiles praised companies working on "access to advanced models and safety testing," suggesting the administration wants to support the development of American AI models as long as it comes with strict security oversight. Anthropic, for its part, stressed that over the two weeks it coordinated with the government to "review and clear" both models for continued availability.

Impact on users and the market

The restored access means businesses and individual users on Pro, Max, Team, and select enterprise plans have regained full functionality of Fable 5 on the Claude platform, in the Claude.AI app, and in Claude Code. The two-week disruption hit non-US business customers particularly hard, forcing many to switch to older models or competing products in the meantime.

The episode shows just how much weight administrative decisions now carry for the global availability of AI models. The sudden export ban could have effectively cut off millions of users outside the United States, including companies using Claude for coding and analytical tasks, from the latest technology. For Polish companies using Claude Code or the Claude platform, it's a reminder that access to advanced models depends not only on decisions made by the developers themselves, but also on Washington's export policy.

For Anthropic, the episode is further evidence that its dealings with the US administration over AI safety will grow more intense as model capabilities keep advancing. The company has pledged to keep working with regulators on standards for future versions of Mythos and Fable, suggesting similar regulatory interventions could recur with upcoming releases.

Sources: White House Lifts Restrictions On Anthropic's Mythos And Fable AI Models (forbes.com), Trump administration lifts export controls on Anthropic's most powerful AI models, ending bitter standoff (yahoo.com), Trump Just Cleared Anthropic to Go Global, and Google DeepMind Should Be Sweating (aol.com)