TL;DR
The U.S. Commerce Department placed Anthropic’s Claude Fable 5 and Mythos 5 under export controls on June 12, and Anthropic disabled both models worldwide. The security basis is disputed, but the business effect is clear: buyers have now seen that access to a frontier U.S. AI model can be cut off by government order.
The U.S. Commerce Department placed Anthropic’s Claude Fable 5 and Mythos 5 under export controls on June 12, forcing the company to disable its newest frontier models for customers worldwide after the order barred access by any foreign national, including some employees, according to Anthropic and reports by Axios and The Verge. The move matters because it turned model access from a product promise into a government-controlled national-security question.
Commerce Secretary Howard Lutnick sent Anthropic CEO Dario Amodei a letter at 5:21 p.m. ET on June 12 placing the two models under export controls, according to the source material. Anthropic said the order covered foreign nationals anywhere, including inside the company, and that it had no clean way to limit access by nationality while keeping the models live for other customers. By midnight, Fable 5 and Mythos 5 were disabled globally.
Fable 5 had been released publicly on June 9 as the guarded commercial version of Anthropic’s Mythos-class system. Mythos 5, the more powerful underlying model, was routed to selected organizations for cyber-defense work through Project Glasswing rather than released broadly.
- Confirmed: The export-control order was issued June 12, covered Fable 5 and Mythos 5, and led Anthropic to shut off both models worldwide.
- Claimed or reported: The government concern centered on alleged jailbreaks, cyberattack-usable outputs and possible access by a China-linked group, according to accounts attributed to the U.K. AI Safety Institute, The Wall Street Journal, Semafor and other outlets.
- Still unsettled: Anthropic disputes that the reported jailbreak justified a global shutdown and says the issue was narrow rather than universal.
Washington just switched off
a frontier model
On June 12, an export-control order forced Anthropic to disable Claude Fable 5 and Mythos 5 worldwide. The security merits are still contested. The lesson buyers took away is not: frontier AI can be turned off.
■ The government’s case
- A reported jailbreak pulled malicious, agentic outputs (UK AISI)
- Amazon told officials Fable yielded cyberattack-usable info
- Suspicion a China-linked group obtained the model
- Proliferation & reverse-engineering risk to national security
▲ Anthropic & 120+ experts
- Calls it a narrow, non-universal jailbreak — a “misunderstanding”
- Capability is real but not unique (GPT-5.5, Opus, Kimi 2.7)
- Controls remove tools from defenders, not just attackers
- Export rules built for chips & ore don’t fit software
The precedent is the story. Whatever the jailbreak’s true severity, the U.S. showed it can dark a commercial American model worldwide on ~90 minutes’ notice. Adoption was supposed to be the moat — this week it became the exposure, and the likely winner is the open, sovereign, self-hosted stack.
Model Access Becomes Contract Risk
The immediate cost to the AI industry is not only lost usage of two models. The larger cost is trust. Enterprise customers, governments and developers now have a live example of a U.S. frontier model being removed from service because of a national-security order. That adds a new risk to long-term contracts, workflow migrations and product plans built around proprietary cloud models.
Axios reported that analysts see the order as a warning for companies whose valuations depend on global adoption of high-end models. Deutsche Bank analyst Jim Reid was cited as warning that if the restriction lasts, it could slow AI integration into global markets. The timing also lands near expected public-market moves by major AI labs, making regulatory shutdown risk harder for investors to ignore.
The shutdown may also push buyers toward multi-model setups, sovereign AI programs and open-weight systems that can be self-hosted. That shift could benefit non-U.S. providers and open models, including Chinese systems, even if U.S. officials intended the controls to reduce national-security exposure.
AI Engineering: Building Applications with Foundation Models
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
A Three-Day Product Reversal
Anthropic launched the Mythos-class models on June 9, presenting them as frontier systems suited for cybersecurity and biomedical work. Three days later, they became the subject of an export-control fight. The source material says the company publicly called the order a misunderstanding and argued that Fable 5 had already gone through thousands of hours of red-teaming by internal teams, the U.S. government, the U.K. AI Safety Institute and outside testers.
The government’s side has not been fully spelled out in a single public technical record. The source material cites a U.K. AI Safety Institute red-team claim that testers found a jailbreak within hours and later extended it to multi-step agentic tool calls. The Wall Street Journal was cited as reporting that Amazon warned officials its researchers obtained information usable in cyberattacks. Semafor was cited as reporting suspicion that a China-linked group had obtained access to the model.
“misunderstanding”
— Anthropic, in its June 12 statement
Access Tools RCBMHD Heavy Duty Button Master
Item model number : ?RCBMHD
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
The Security Case Remains Disputed
It is not yet clear whether the reported jailbreak represented an exceptional danger, a patchable safety failure or a capability already common across frontier and open-weight models. It is also unclear whether a China-linked group obtained the model, whether Amazon’s reported findings were decisive in the order, or what technical fix would satisfy the government. Anthropic says access to its other models was not affected, but the duration of the Fable 5 and Mythos 5 restrictions remains unresolved.
Artificial Intelligence for Cybersecurity: Develop AI approaches to solve cybersecurity problems in your organization
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
June 22 Talks Awaited
Anthropic and White House officials are expected to meet on June 22. The key questions are whether Commerce lifts or narrows the controls, whether Anthropic offers a patch or access structure that satisfies officials, and whether the U.S. creates a broader export-control framework for frontier software models. Customers and investors will be watching less for rhetoric than for proof that advanced AI services can remain available after launch.
AI Model Risk Blueprint: Model Validation Testing | Ethical Considerations in AI Models | Integrating AI with Business Risk Plans | Real-World AI Model … Strategies | AI Governance Tools & Resource
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Key Questions
What exactly did the U.S. government do to Anthropic?
The Commerce Department placed Claude Fable 5 and Mythos 5 under export controls on June 12, barring access by foreign nationals. Anthropic said it could not comply selectively, so it disabled both models worldwide.
Why were the models restricted?
The stated basis was national security. Reports cited concerns about jailbreaks, cyberattack-usable outputs and possible foreign access. Anthropic disputes that the known evidence justified recalling the models.
Does this affect all Claude models?
No. According to the source material and related reporting, the order covered Claude Fable 5 and Mythos 5. Anthropic said access to its other models was not changed.
Why does this matter beyond Anthropic?
The shutdown gives companies a concrete reason to treat proprietary frontier AI as a regulated supply chain, not just a cloud service. That may push customers to diversify providers, keep backup models, or favor self-hosted systems.
When could the ban change?
The next known milestone is a scheduled June 22 meeting between Anthropic and the White House. Any change would depend on government approval, technical fixes or a narrowed access policy.
Source: Thorsten Meyer AI
