Anthropic has suspended all customer access to its Fable 5 and Mythos 5 AI models to comply with a US government directive issued on June 12. The Commerce Department cited national security concerns tied to a potential jailbreak. Other Anthropic models remain available.
The company acted after receiving verbal evidence from officials about a narrow jailbreak that could allow Fable 5 to review code for software flaws. Anthropic said the evidence involved only minor vulnerabilities and noted that similar capabilities exist in other public models.
Fable 5 launched on June 9 as a public version of capabilities previously limited to Mythos 5, which is reserved for Project Glasswing partners. The government order imposes export controls that restrict use outside the United States.
Anthropic said it disagrees with the decision to recall the models but is complying fully. It plans to share more details within 24 hours and called for a transparent statutory process for such actions.
President Trump signed an executive order earlier this month that encourages voluntary government security testing for new AI models.
