Anthropic has suspended access to its Fable 5 and Mythos 5 artificial intelligence models following a direct order from the US government. The directive, received at 5:21 pm ET on Friday, mandates an immediate halt to all access by foreign nationals, including Anthropic’s own international staff.
The company disabled the models abruptly to ensure full compliance with the legal instruction. Anthropic confirmed that other offerings, such as Opus 4.8, remain operational and unaffected by this restriction.
These models were released just days prior, built upon the Mythos Preview framework which demonstrated significant capabilities in identifying software vulnerabilities. However, federal authorities raised concerns regarding a specific method capable of bypassing Fable 5's safety protocols.
Anthropic stated that officials provided only verbal evidence of a narrow, non-universal jailbreak technique involving code analysis. The firm argued this poses a significantly lower threat than a universal bypass and warned that applying such strict standards industry-wide could effectively stall innovation in frontier AI development.
Anthropic maintains that the directive stems from a misunderstanding of the risk level. The company is actively engaging with authorities to clarify the situation and restore user access as quickly as possible.