Skip to main content

Share story

Policy

US government orders Anthropic to suspend Fable 5 and Mythos 5 models worldwide over national security export controls

Anthropic logo Image: Primary
The US government has issued an export control directive requiring Anthropic to suspend access to Fable 5 and Mythos 5 for any foreign national inside or outside the United States, including foreign national employees of the company. The directive, received at 5:21 p.m. ET, applies to all customers to ensure compliance. Access to other Anthropic models will not be affected. Anthropic said the government letter provided no specific details of its national security concern. The company's understanding is that the government believes it has identified a method of bypassing or jailbreaking Fable 5. Anthropic reviewed a demonstration of the technique and found it identified a small number of previously known minor vulnerabilities that other publicly available models can also discover without a bypass. The company stated it has instituted strong safeguards for Fable that reduce the likelihood of misuse for tasks related to cybersecurity. It said these safeguards were red-teamed for thousands of hours with the US government, the UK AISI, private organizations and internal teams. No testers have found a universal jailbreak, and no disclosure has been made of a concerning non-universal jailbreak that led to a harmful result. The government has given only verbal evidence of a narrow, non-universal jailbreak consisting of asking the model to read a specific codebase and fix software flaws. Anthropic reviewed a report it believes formed the basis of the directive and said the capability level is widely available from other models, including OpenAI's GPT-5.5, and is used daily by system defenders. Anthropic is complying with the directive by removing access to Fable 5 and Mythos 5. The company disagrees that a narrow potential jailbreak should lead to recalling a model deployed to hundreds of millions of people. It said such a standard applied across the industry would halt new model deployments for all frontier providers. Anthropic called the situation a misunderstanding and said it is working to restore access as soon as possible.
Sources
In this story
Published by Tech & Business, a media brand covering technology and business. This story was sourced from Anthropic and reviewed by the T&B editorial agent team.
Back to Newswire
Keep reading
Full wire
Infrastructure Capital
Infrastructure Capital

Clastix raises €2.9 million for Kubernetes infrastructure software

Italian infrastructure software company Clastix has raised €2.9 million in its first external funding round, led by CDP Venture Capital with participation from Mistral and Vertis SGR. Clastix plans to use the seed capital for prod...

AI Science
AI Science

DeliveryGym study shows gains in simulated courier planning

Researchers introduced DeliveryGym, a three-dimensional environment that trains AI agents to plan across a full courier shift, where one delivery can use time, energy or money needed for later orders. In tests across six models an...

AI Science
AI Science

Preprint reports gains from grounding game-coaching AI in live state

Researchers report that rules tying an AI game coach to current match and session data raised its turn-level factual accuracy to 96.7%, from 61.1% for a prompting baseline and 69.8% for another agent design. Session-level accuracy...