US Government Orders Suspension of Anthropic’s Fable 5 and Mythos 5 Models

Anthropic was directed by the US administration to immediately suspend all access to its two most advanced AI models, Fable 5 and Mythos 5. The Government cited national security concerns.

Last week’s directive told the AI lab to disable both models for all customers globally, including foreign national Anthropic employees, whether inside or outside the US. The Trump administration’s letter to Anthropic targeted foreign nationals, but because ensuring compliance would require verifying the nationality of every user, it shut both models down entirely for all customers, Anthropic stated in a blog.

The letter did not specify the nature of its national security concern, but Anthropic said its understanding is that government officials believe a method of bypassing the model’s safeguards through jailbreaking has been identified. The AI player stated it reviewed a demonstration of the technique and found it exposed only a small number of previously known minor vulnerabilities, all of which can be replicated using other publicly available AI models without any bypass required.

Anthropic pushed back on the administration’s justification for the suspension. The company noted that in the weeks building up to Fable 5’s launch, it worked extensively with the US government, the UK AI Safety Institute, and multiple private organisations. Those tests, Anthropic said, showed Fable 5’s protections were substantially stronger than any previously deployed model. No tester has yet found a universal jailbreak capable of broadly bypassing the model’s defences, according to Anthropic.

It did acknowledge perfect jailbreak resistance is not currently achievable for any AI model provider and stated it disclosed this limitation publicly at launch. Anthropic adopted what it describes as a defence-in-depth strategy, combining narrow jailbreak resistance with thorough monitoring and a 30-day customer data retention policy specifically designed to detect and shut down successful attacks.

The directive followed conversations between Amazon CEO Andy Jassy and US officials including Treasury Secretary Scott Bessent, after the retail giant’s researchers used a series of prompts to extract information from Fable 5 that was restricted. Anthropic said it is complying with the legal order but disagrees that the narrow potential jailbreak warrants recalling a model deployed to hundreds of millions of users. The company warned that applying this standard across the industry would effectively halt all new frontier model deployments. Some customers are seeking refunds from Anthropic for the disabling of Fable 5, citing a post on Reddit from the company which offers a prorated refund. Anthropic apologised for the disruption, describing the situation as a misunderstanding, and said it is working to restore access as soon as possible.