The U.S. Department of Commerce has lifted its export ban on Anthropic’s advanced AI models, Claude Fable 5 and Mythos 5, ending a nearly three-week suspension that raised questions about national security and the reach of American technology controls. Access to Claude Fable 5 will be restored across its main platforms on Wednesday, the company confirmed.
The ban was imposed on June 12 after researchers discovered a potential “jailbreak” – a method to bypass the model’s safety guardrails – that could allow users to request cybersecurity-related outputs, including software vulnerability discovery. Anthropic argued the issue was narrow and not unique to Fable 5, but complied with the directive, which also blocked access for foreign nationals, including some of its own staff.
Following two weeks of review with the Commerce Department, officials approved the redeployment after Anthropic agreed to add new classifiers that detect and block risky cybersecurity requests. When a request is flagged, the system can route the user to a lower-risk model. The company also pledged ongoing collaboration with U.S. agencies on pre-release reviews, jailbreak reporting, and misuse tracking.
The ban’s sudden enforcement drew wider policy attention. Austria later urged the European Union to consider ways to secure access to Anthropic systems, highlighting how U.S. export decisions affect technology users globally. While Fable 5 – a consumer-focused model for deep reasoning – is immediately returning, Mythos 5 remains more limited; only a few approved U.S. organizations have regained access for security work, and broader domestic and international availability awaits further government approval.
The announcement also spotlighted infrastructure partners Amazon and Broadcom. Amazon has invested $13 billion in Anthropic and committed up to $20 billion more, with a $100 billion-plus cloud deal over ten years. Broadcom is set to supply massive computing capacity from 2027. While the stock reaction was muted, the restored access removes a key overhang for both companies. Separately, both Anthropic and OpenAI are reportedly preparing IPOs with valuations exceeding $1 trillion later this year.
