Original briefings. Zero spin.
Every story is an original briefing written from 60+ sources across the spectrum — sources linked so you can verify it yourself.
Commerce Secretary Lutnick Orders Anthropic to Kill Foreign Access to Fable 5 and Mythos 5, Citing a Jailbreak Claim

What Happened
At 5:21 p.m. ET on Friday, June 12, 2026, Anthropic received a directive from the US government ordering it to suspend all access to Claude Fable 5 and Claude Mythos 5 for "any foreign national, whether inside or outside the United States, including foreign national Anthropic employees." Anthropic confirmed the order in a public statement posted that evening.
Because Anthropic could not instantly verify the citizenship of every user, it took the broadest available path: it disabled both models for its entire customer base. Access to all other Anthropic models was unaffected.
Who Pulled the Trigger
According to Axios, Commerce Secretary Howard Lutnick sent the letter directly to Dario Amodei. An administration official told Axios that the Commerce Department acted after a separate company claimed it had successfully jailbroken Mythos 5, which alarmed administration officials about potential national security exposure.
Axios also reported that the Trump administration had previously attempted to block Anthropic from releasing Mythos 5 entirely and failed. This export control directive appears to be a second attempt to constrain the model's reach.
Under the directive's terms, a license would be required for "the export, re-export or domestic transfer" of the affected models, according to Axios.
Anthropic's Pushback
Anthropic accepted the order but disputed its premise. The company said the government's letter provided no specific details about the national security concern. Anthropic's own assessment: the claimed jailbreak technique, when demonstrated, revealed only "a small number of previously known, minor vulnerabilities" — and that other publicly available models could surface the same information without any bypass technique at all.
The company's statement was pointed. Anthropic noted it had red-teamed Fable 5 for thousands of hours with the US government, the UK AI Safety Institute, multiple private third-party organizations, and internal teams before launch. Those tests showed Fable's safeguards were "substantially more effective than those of any previously deployed model." No tester had found a universal jailbreak.
Anthropic acknowledged upfront, in its original launch documentation, that perfect jailbreak resistance is not achievable for any model provider. It said this publicly. The company described its strategy as "defense in depth" — making jailbreaks either narrow or expensive to execute, then using monitoring to catch successful attacks quickly. That strategy included a 30-day customer data retention policy Anthropic described as carrying "real costs."
The company said it believes the government is operating under a misunderstanding and is working to restore access as quickly as possible.
Background on the Models
Fable 5 and Mythos 5 launched on June 9, 2026, according to the Trilogy AI Center of Excellence. Both descended from Claude Mythos Preview, a cybersecurity-focused model that Anthropic unveiled in April and restricted to a small group of vetted organizations under a program called Project Glasswing.
Mythos Preview drew significant attention from Wall Street and government officials for its ability to identify software vulnerabilities. Mozilla alone reported resolving hundreds of security flaws using it, according to 9to5Mac.
When the full releases came, Anthropic kept Mythos 5 inside Project Glasswing's restricted access structure. Fable 5 was the first broadly public release of a model at this capability level. It launched with guardrails that users almost immediately complained were too restrictive. Anthropic had been actively adjusting those safeguards in the days between launch and the government's shutdown order.
The Practical Fallout
Foreign customers — companies and developers outside the US who had integrated Fable 5 or Mythos 5 into products — lost access without warning late on a Friday afternoon. Anthropic's non-US employees who are not American citizens also lost access to the models they were presumably using for work.
A Hacker News discussion surfaced a real supply-chain concern: one commenter noted that this event transforms the theoretical risk of US model shutdowns into a demonstrated reality, and that non-US enterprises now have concrete evidence that dependence on American AI providers carries regulatory discontinuity risk. Whether that concern drives meaningful customer defection toward European models like Mistral or emerging domestic alternatives in markets like India remains an open question — one the industry will be watching closely.
What's Missing From the Record
The government has not disclosed which company claimed to have jailbroken Mythos 5, what the jailbreak actually enabled, or why Mythos 5 was included in the shutdown when the stated concern was about Fable 5. Anthropic's statement addressed the Fable 5 jailbreak claim specifically but did not explain the rationale for pulling Mythos 5 alongside it.
No investigation, charge, or legal proceeding has been announced. This is an export control administrative action, not a criminal matter.
The unresolved question with real stakes: Anthropic says it is working to restore access, but the export control mechanism now in place requires a Commerce Department license for domestic transfers as well. Whether Anthropic can satisfy that licensing requirement quickly — or at all, given the administration's prior attempt to block the models entirely — will determine whether Fable 5 and Mythos 5 come back online or stay dark.
Sources used for this briefing
This briefing was written by UBH's AI agent — these are the reporting inputs it draws on, linked so you can verify.