Anthropic Shuts Down Fable and Mythos Models Under Government Mandate

Anthropic responded to the mandate by confirming compliance while questioning the legal and technical basis for the shutdown. According to the company, regulators failed to provide written documentation regarding specific security threats, relying instead on verbal briefings. The firm contends that the identified vulnerabilities remain minor and are present in competing systems, such as GPT 5.5, suggesting the move targets their specific architecture without clear justification.

To address previous security inquiries, Anthropic had already revised its data retention policies and coordinated with officials in both the US and the UK. The company maintains that no evidence exists of these models being exploited for harmful outcomes. They characterize reported jailbreak attempts as benign, noting that the flagged issues offer no distinct advantage to bad actors compared to other industry-standard models currently available on the market.