U.S. Export Controls Halt Anthropic’s Fable 5 and Mythos 5 AI Models

On June 12, Commerce Secretary Howard Lutnick mailed Anthropic CEO Dario Amodei a letter invoking the 2018 Export Control Reform Act, ordering suspension of the Claude Fable 5 and Mythos 5 models and demanding a license for any export or transfer to foreign nationals, warning of “prompt criminal and civil penalties.” Anthropic disabled the models for all users on June 13.

UsNews and World Report

The conflict traces to early 2024 when the Pentagon labeled Anthropic a supply-chain risk after the firm refused to drop contractual limits barring its AI’s use in lethal autonomous weapons and domestic surveillance. The refusal led to a national-security blacklist and informed the Trump administration’s National Security Presidential Memorandum 11, which allows termination of contracts with AI providers that restrict government use.

What Happens When the US Military Puts AI in Charge of National Security?

Over 80 cybersecurity executives and a coalition of 42 CISOs from firms NVIDIA, Adobe, Zoom and Veracode signed open letters urging the ban be lifted, arguing that similar vulnerability-finding abilities exist in other models (e.g., GPT-5.5, Opus, Kimi 2.7) and that the restriction “unilaterally disarms Western cyber defenders” without hindering adversaries.

Security leaders say lift export controls for Anthropic's Mythos-class models

Lutnick warned that China, Russia or other countries of concern could exploit the models, invoking ECRA authority and threatening “prompt criminal and civil penalties.” Anthropic said the concern stemmed from a narrow, non-universal jailbreak that revealed only minor flaws already present in public models, and noted Fable 5’s built-in safeguards. The Commerce Department has not clarified how export controls apply to cloud-based AI.

Anthropic pulls Mythos-class models globally

Export-control scholars note AI models are accessed remotely, not shipped, questioning the 2018 Export Control Reform Act’s relevance. The government cites a potential jailbreak; industry experts deem the vulnerability minor and already replicated. No detailed technical report has been released.

It turned out that the reason why the U.S. government completely suspended the latest models of arti. .

• “There is a risk that military and intelligence agencies in countries of concern such as China and Russia will use the model,” — Howard Lutnick, U.S. Commerce Secretary

It turned out that the reason why the U.S. government completely suspended the latest models of arti. .

• “Many of the undersigned individuals regularly use other foundation and open-source models for security audits and red-teaming every day,” — Open letter from security leaders

Security leaders say lift export controls for Anthropic's Mythos-class models

• “The unintended consequence of restricting access to the Mythos-class models is that it will harm defence while doing nothing to impede attackers, Moussouris said.” — Katie Moussouris, Luta Security chief executive

Security leaders say lift export controls for Anthropic's Mythos-class models

Anthropic’s senior technical team holds daily virtual talks with Commerce officials; both Lutnick and CEO Dario Amodei will attend the G7 summit in Evian-les-Bains, where the issue may be discussed. The department signaled a license could be granted after confirming the models cannot be “jailbroken,” while industry groups seek legislative clarification of export-control authority over cloud-based AI.

UsNews and World Report