On June 12, 2026, the US Commerce Department issued a national security export-control directive to Anthropic requiring it to suspend access to the AI models "Fable 5" and "Mythos 5" for foreign nationals. Anthropic responded by disabling both models for all users worldwide.
June 12, 2026 · Anthropic
US Export Order Forces Anthropic to Shut Down Its Two Most Advanced Models
A Commerce Department directive barred foreign-national access to Claude Fable 5 and Mythos 5. Unable to check nationality through its API, Anthropic suspended both models for all users immediately.
2
advanced models suspended (Fable 5 & Mythos 5)
95%
of synthetic jailbreaks blocked by classifiers
76
security experts signed a letter demanding reversal
1,000+
hours of bounty work found no universal jailbreak
Which models are affected
Claude Fable 5 Public-facing
SUSPENDED
Claude Mythos 5 Advanced base model
SUSPENDED
Opus 4.8 / Sonnet / Haiku Mythos traffic rerouted to Opus 4.8
AVAILABLE
The "jailbreak" at the center of it all
Experts argue the flagged technique amounts to little more than rephrasing the same routine security request — a capability defenders use every day.
Allowed
"Review code for security issues"
→
Support
Some customers praised Anthropic's safety-first posture and willingness to comply fully rather than skirt the directive.
Pushback
Security veterans call it a dangerous precedent that strips defenders of key tools — with no transparent legal process or technical evidence.
First of its kind: the directive extends US export controls to AI models themselves — beyond chip rules — even as rivals are said to possess similar cyber capabilities. Open questions remain on the legal process and any timeline for restoration.
Continue reading The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.
Already purchased? Sign in ✓ Signed in — this article isn’t included in your current plan.Unlocking the full article…