US government orders Anthropic to restrict access to Fable 5 and Mythos 5
The US government issued an export control directive on June 12, 2026, requiring Anthropic to ban non-Americans, including some company employees, from accessing its Fable 5 and Mythos 5 AI models. Anthropic immediately suspended both models from the market, reportedly over concerns about a jailbreak that could bypass their safeguards, though the exact cause was not specified by the government.
What happened
On June 12, 2026, the US government issued an export control directive to Anthropic requiring the company to suspend all access to Fable 5 and Mythos 5 by foreign nationals, including some Anthropic employees.
Anthropic immediately pulled both models from access, citing the difficulty of auditing the citizenship of every user. The company reportedly did not receive a specific reason for the ban but believes it is related to a recently discovered jailbreak that bypasses the models' safeguards.
Reported impact
Anthropic was forced to withdraw its most powerful models from the market shortly after their launch, preventing public access to both Fable 5 and Mythos 5.
Index analysis
Fable 5 was created by adding safeguards to the Mythos model to make it suitable for public release. Anthropic had marketed Mythos as dangerous and restricted its public release because the model is so capable at finding security flaws in software code.
Governance controls
Anthropic argues that Fable 5's safeguards effectively prevent its use for malicious purposes and claims that the outlined vulnerabilities are neither new nor significant. The company also maintains that other publicly available models can discover the same vulnerabilities without jailbreaking.
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.