United States lifts export controls on Anthropic models Fable and Mythos

The US government has lifted export controls on Anthropic's most advanced AI models, Claude Fable 5 and Mythos 5, and will begin restoring access from tomorrow, the company announced late Tuesday. The decision removes barriers that had blocked foreign nationals and some organizations from using the models and arrives as Washington intensifies oversight of frontier AI releases.

Export controls lifted after security agreement

Anthropic abruptly restricted access to Claude Fable 5 and Mythos 5 last month after the Trump administration ordered the company to bar all foreign nationals, including its own employees, from the models. The government cited unspecified national security concerns, though reports circulated about vulnerabilities in Fable 5 that allowed researchers to "jailbreak" the system.

Commerce Secretary Howard Lutnick informed Anthropic that it would no longer need an export licence. In a letter widely shared online, Lutnick said the company had agreed to "proactively detect and address security risks associated with the models," work with the government on standards for upcoming releases, and report malicious activity. Anthropic said on X, "We're grateful to our users for their patience, and to everyone who worked with us on redeploying the models."

The restoration first extends to US organizations that "operate and defend critical infrastructure," a permission granted last Friday, before rolling out more broadly. For IT professionals working with these systems, Claude AI Courses offer practical training on Anthropic's tools and their deployment.

Anthropic's strained relationship with the administration

The export restrictions escalated tensions between Anthropic and the Trump White House. In March, the company sued the Department of Defense after the Pentagon labelled it a "supply chain risk." Anthropic had refused to work with the military without explicit assurances that its AI would not be used for mass surveillance or autonomous weapons.

While Trump started his second term championing a laissez-faire approach to AI regulation, his administration has since tightened scrutiny. OpenAI announced last week that it would release its latest model series, GPT-5.6, first to a "small group of trusted partners" following government pressure to stagger access.

Industry experts question regulatory precedent

Francesco Bailo, deputy director of the AI, Trust and Governance Centre at the University of Sydney, said reports of jailbreaking Fable 5 had been exaggerated. "The US government likely realised it had overreacted, and also that its decision would produce a dangerous, messy precedent in terms of regulations and strong backlash from an industry that has invested considerably in maintaining close communication with the Trump administration," he said. Bailo added that "if Fable and Mythos were blocked on these grounds, competitor models would have to be blocked too."

Tanishq Abraham, a former research director at Stability AI who now leads medical AI company Sophont, said the lift is a "big deal" that raises important regulatory questions. "Externally, it seems like the fraught relationship between Anthropic and the US government has been improving, and this is likely due to the efforts of Tom Brown," he said, referring to the Anthropic cofounder. Abraham pointed to the broader uncertainty: "The biggest question now is: What precedent does this set for the industry? Does the US government need to approve every frontier model release? We're already starting to see this with the GPT-5.6 release. Lots of unanswered questions about how frontier labs have to interact with the US government."

Why this matters for IT, development, and government professionals

For IT and development teams, the restoration of Claude Fable 5 and Mythos 5 means immediate access to some of the most capable AI models on the market, particularly for tasks in critical infrastructure and secure environments. The rapid policy shifts, however, signal that access to such tools can change quickly, making contingency planning essential.

Government professionals face a fast-evolving regulatory landscape where export controls on AI are becoming a standard tool. Understanding the frameworks that govern model releases - and the security agreements that can unlock them - will directly shape procurement and compliance decisions. AI for Government training can help navigate these emerging requirements. The case also strengthens the argument that close coordination between agencies and frontier labs is now a permanent fixture of AI deployment.