US export controls hit top Claude models
The US ordered Anthropic to suspend Fable 5 and Mythos 5; the company cut access for all users worldwide, including its own staff, and is disputing it.
On June 13, Anthropic confirmed it had suspended its two most powerful models, Fable 5 and Mythos 5, for every user worldwide, including US customers and its own staff, after the US government ordered it to. The directive reached the company at 5:21pm ET on June 12, the same Commerce Department action Axios first reported as export controls on foreign access.
Anthropic says the government believes someone found a way to jailbreak Fable 5 by asking it to read a codebase and fix software flaws, and that the demonstrations it saw surfaced only a few previously known, minor vulnerabilities. The company disputes the call, arguing the same capability ships in rival models, including OpenAI’s GPT-5.5, and that cybersecurity teams rely on it daily:
If this standard was applied across the industry, we believe it would essentially halt all new model deployments for all frontier model providers.
It is complying while it works to restore access, with more detail promised within 24 hours.
Why this matters
What began as a reported foreign-access limit is now a full shutdown of a paid product teams had built into real workflows, and it reaches US companies too. Export controls were built for chips and weapons parts; pointed at a subscription model, they mean your access can vanish on a government letter rather than a billing dispute.
The defensive move is the boring one that always works: know which model your workflows depend on, and rehearse the swap before you need it. Our guide to vendor lock-in exit ramps keeps that switch cheap, and how to choose AI tools covers weighing a vendor’s stability before you commit. It lands the same week Anthropic changed how Fable 5’s guardrails behave, so the ground under this vendor is moving on two fronts at once.
Update: the EU pushes back
On June 14 the European Commission said it is studying what the US order means for European users of Claude, and that any contingency measures “should not be discriminatory against partners.” Spokesperson Thomas Regnier framed it as proof Europe needs to strengthen its own technological sovereignty, noting that capable new models help cyber-defence even as they raise security concerns. No countermeasure was announced. For operators outside the US, the takeaway is unchanged but sharper: access to a foreign vendor’s flagship model can now turn on decisions made in another government’s capital, so the fallback you rehearsed is the part you actually control.
Update: still offline, and the rift runs deeper
Four days on, Fable 5 and Mythos 5 remain suspended with no restoration date, and Anthropic and the White House are still at odds. Anthropic maintains the evidence against it is thin:
We have not even received a disclosure of a concerning non-universal potential jailbreak that led to a harmful result.
Reporting from the Washington Post traces the break to a longer fight: the administration had already labeled Anthropic a “supply chain risk” after the company refused to drop limits on using Claude for mass surveillance and autonomous weapons. The practical line for operators only hardens. A flagship you rent can stay dark for days over a dispute you are not party to, so the fallback you rehearsed is still the part you control.