Gulf users among those cut off as Anthropic disables Claude Fable 5 on US order

Dubai, UAE: Users in the UAE and across the Gulf are among those who have lost access to Anthropic’s most advanced AI model, after the US government ordered the company to bar foreign nationals from using it.

Anthropic said on Friday it would disable Claude Fable 5 and Claude Mythos 5 for all customers, after a US export-control directive suspended access for foreign nationals on national security grounds, according to a company statement reported by Reuters. Because the order applies worldwide, users in the region fall within its scope.

The company said the directive arrived at 5.21pm Eastern time on Friday and reaches foreign nationals inside and outside the United States, including Anthropic’s own foreign-born staff, the South China Morning Post reported. Because the order could not be applied selectively across a shared cloud service, the company said it had to switch off both models for everyone.

Anthropic said the “net effect” of the order was that it had to disable both models for all customers to comply. It added that access to its other models would not be affected.

What the US export order on Anthropic covers

Commerce Secretary Howard Lutnick sent a letter to Anthropic chief executive Dario Amodei outlining the restrictions, according to Axios and The Wall Street Journal. NBC News reported that the letter was prepared with help from officials at the Commerce Department’s Bureau of Industry and Security, and that the department did not immediately respond to a request for comment. Anthropic said the letter gave no specific detail of the national security concern behind it.

Anthropic said its understanding is that the government believes a safeguard in Fable 5 can be bypassed, or “jailbroken.” That safeguard is meant to stop the model being used to identify software vulnerabilities.

An administration official told Axios that the Commerce Department acted after another company claimed it had jailbroken Mythos. The official said this raised concern about possible national security risks. The same official said the administration had earlier asked Anthropic to delay the launch, and that the company declined.

Only Fable 5 and Mythos 5 are affected. Anthropic said access to its other models is unchanged.

As the change took effect, users who had access reported seeing the message “Claude Fable 5 is currently unavailable” in place of the model.

Claude Fable 5 and Mythos 5: how the two models differ

Anthropic released Fable 5 and Mythos 5 earlier in the week, describing them as its most capable models to date and as part of a new tier it calls “Mythos-class.” The two share a similar technical foundation. Fable 5 was released to the public with added safety and security restrictions, including limits on responses about cybersecurity, biology and chemistry. Mythos 5 was never offered to the public. It was made available only to a vetted set of partners under a programme the company calls Project Glasswing.

That difference shapes what the shutdown means in practice. For most users, including those in the Middle East, the change is the loss of Fable 5, the public model. Mythos 5 was already out of reach for general users before the order.

What the Fable 5 shutdown means for users in the UAE and Gulf

The directive applies to all foreign nationals worldwide, so users, developers and businesses in the UAE and across the Gulf fall within its scope. Anyone in the region who had started using or testing Fable 5 since its launch this week would have lost access as the order took effect.

For now the practical effect in the region is limited to these two models. Anthropic said its other models, which many regional users and companies already rely on, remain available. The company has not said how long the suspension will last, so businesses in the Gulf that were planning around Fable 5 face an open question on timing.

How Anthropic responded to the US order

Anthropic said it disagrees with the rationale but is complying with the directive. The company said it believes the order rests on a misunderstanding and that it is working to restore access as soon as possible, Business Standard reported, though it did not give a timetable.

The company said it had reviewed what it believes is the report behind the government’s action. It described the capability shown as a narrow jailbreak, not a universal one. It said similar capability is available in other public models, including OpenAI’s GPT-5.5, and that cybersecurity teams use such techniques routinely.

Anthropic also said the models went through thousands of hours of red-teaming before launch. That testing included the US government and the UK AI Security Institute. None found a universal jailbreak, the company said.

Why the timing matters: Anthropic’s IPO and Washington tensions

The shutdown lands at a delicate time for the company. Anthropic confirmed on 1 June that it had confidentially filed a draft prospectus with the SEC for a proposed initial public offering, a step also reported by CNBC and Quartz. This is not its first clash with Washington. Earlier this year the Department of Defense designated the company a supply chain risk, a label Anthropic is challenging in court.

Claude Fable 5 and the US order: frequently asked questions

Which Anthropic models have been disabled?

Only Claude Fable 5 and Claude Mythos 5. The company said all of its other models remain available.

Why did Anthropic disable them?

The company said it received a US export-control directive suspending access for foreign nationals, and that it could not enforce a nationality-based limit selectively, so it switched the two models off for everyone to comply.

Who issued the order?

Commerce Secretary Howard Lutnick sent a letter to chief executive Dario Amodei, according to Axios and the Wall Street Journal. The directive did not set out the specific national security concern, Anthropic said.

What is the security concern?

Anthropic said its understanding is that the government believes a safeguard in Fable 5 could be bypassed to help identify software vulnerabilities. The company said the capability is narrow and is available in other public models.

Will access return?

Anthropic said it believes the order is based on a misunderstanding and that it is working to restore access. It did not give a timetable.