Claude Fable 5 Returns Worldwide as US Lifts Its 18-Day Export Control Order

On July 1, 2026, Anthropic restored Claude Fable 5 to full public availability worldwide. The model had been offline for foreign users since June 12, after the US government issued an export control directive ordering Anthropic to cut off access for anyone who wasn’t a US national, wherever they were located. Because Anthropic had no reliable way to check nationality at the point of a chat request, the company ended up disabling Fable 5 for everyone, US customers included.

Eighteen days later, the order was reversed. The Commerce Department lifted the restriction on June 30, and Fable 5 came back online globally the next day.

TL;DR

Anthropic launched Claude Fable 5 on June 9, 2026. Three days later, the US government issued an export control directive requiring Anthropic to block all foreign nationals from the model, citing a jailbreak technique that Amazon’s security researchers had surfaced during testing. Anthropic couldn’t verify nationality in real time, so it pulled Fable 5 (and the related Mythos 5) for every customer worldwide rather than just foreign accounts. Anthropic disagreed with the government’s read on the risk but complied. Over the following weeks, Anthropic built a new safety classifier that blocks the specific bypass technique in more than 99% of attempts and reroutes flagged requests to Claude Opus 4.8. Commerce reviewed the fix, lifted the export controls on June 30, and Anthropic switched Fable 5 back on globally on July 1, across Claude.ai, the Claude Platform, Claude Code, and Claude Cowork, with cloud access on AWS, Google Cloud, and Microsoft Foundry following.

What the export control order actually said

The directive arrived on June 12, 2026 at 5:21 PM ET. Anthropic’s own account of it is specific: the order required the company to “suspend all access to Fable 5 and Mythos 5 by any foreign national, whether inside or outside the United States,” including Anthropic’s own foreign national employees.

That’s an unusual shape for an export control. It didn’t name a country. It drew a line around a class of person, defined by the passport they hold, applied globally.

The trigger was a jailbreak report. Amazon security researchers, during testing, found a way to get Fable 5 to help identify software vulnerabilities by asking it to read a specific codebase and flag flaws. In one case the model produced code demonstrating how to exploit one of those flaws. The government’s read was that this crossed a national security line significant enough to justify pulling access.

Anthropic’s read was different, and the company said so publicly. Its statement called the vulnerabilities “relatively simple” and noted that other publicly available models, including Claude Opus 4.8, GPT-5.5, and Kimi K2.7, were able to find and demonstrate the same exploits when tested the same way. In other words: nothing about the flaw was unique to Fable 5’s added capability.

Anthropic complied with the order anyway, while disputing the reasoning on the record: “we disagree that the finding of a narrow potential jailbreak should be cause for recalling a commercial model deployed to hundreds of millions of people.”

What changed between June 12 and July 1

Anthropic spent the suspension building a fix rather than fighting the order in public.

• A new safety classifier trained specifically to detect and block the reported bypass technique, which Anthropic says now catches the pattern in over 99% of attempts.
• Requests that trip the classifier get automatically rerouted to Claude Opus 4.8 instead of being served by Fable 5.
• Anthropic says it doubled its safety research staffing in the weeks around the launch.
• Commerce evaluated the mitigation and, on June 30, lifted the export controls on both Fable 5 and Mythos 5.

The government’s own statements on the reversal frame it as a deliberate, negotiated close rather than a quiet walk-back. Commerce Secretary Howard Lutnick said the department had “worked closely” with Anthropic to review and approve Fable 5’s return, and credited the company for coordinating to “proactively detect and address security risks.” White House Chief of Staff Susie Wiles framed the priority as getting advanced AI “deployed as quickly and safely as possible.”

Commerce also kept a hook in the deal. It reserves the authority to reimpose restrictions if circumstances change or if Anthropic falls short on the malicious-activity detection commitments it made.

What “full public availability worldwide” means in practice

As of July 1, Fable 5 is back for every user, foreign national or not, across:

• Claude.ai and the Claude Platform (Anthropic’s consumer and developer-facing surfaces)
• Claude Code
• Claude Cowork
• AWS, Google Cloud, and Microsoft Foundry, where Anthropic is re-enabling cloud access on a rolling basis

Anthropic also flagged a temporary usage note for the relaunch: Pro, Max, Team, and select Enterprise plans got a 50% usage allocation bump through July 7, likely a ramp-up measure while the reactivated infrastructure absorbed demand that had been redirected to other models during the pause.

Mythos 5, the related model caught in the same order, was restored on the same timeline.

Why this is a bigger story than one model outage

Export controls on AI models aren’t new. Export controls that target a technique rather than a country, applied globally and retroactively to a model already in the hands of hundreds of millions of users, are new.

That’s the part worth sitting with. The order didn’t restrict Fable 5 from shipping to a new market. It reached back into a live, already-deployed product and shut it off worldwide, because Anthropic had no way to comply with a nationality-based rule at the level of an individual chat request. The compliance mechanism didn’t exist, so the whole model went dark instead.

It also put Anthropic in the position of contradicting the government’s technical judgment in public while still following the order. The jailbreak was real, but Anthropic’s own testing said it wasn’t unique to Fable 5. The government pulled the model for three weeks anyway. Whether that was the right call is above this site’s pay grade. What’s clear is the precedent: a US administration will yank a released frontier model over a security finding, then negotiate a return once a fix is built and reviewed.

There’s a narrower angle worth naming for anyone tracking AI access generally. An 18-day gap where a major model was simply unreachable for most of the world is a live demonstration that a government order is now a real failure mode for anything built on a specific frontier model, not a hypothetical one.

Frequently asked questions

When was Claude Fable 5 restored to full availability? July 1, 2026. The US Commerce Department lifted the export controls on June 30, and Anthropic brought Fable 5 back online globally the next day.

Why was Claude Fable 5 restricted in the first place? On June 12, 2026, the US government issued an export control directive ordering Anthropic to block foreign nationals from accessing Fable 5 and Mythos 5, citing a jailbreak technique Amazon’s security researchers found during testing. Anthropic couldn’t enforce a nationality check in real time, so it disabled both models for all users worldwide.

Was the jailbreak a real security risk? Anthropic disputed how serious it was. The company said the vulnerabilities the technique surfaced were relatively simple and that other public models, including Opus 4.8, GPT-5.5, and Kimi K2.7, could find the same ones. The US government still treated it as significant enough to justify the export controls.

What did Anthropic change to get the restrictions lifted? A new safety classifier trained to block the specific bypass technique, which Anthropic says now works in over 99% of attempts, with flagged requests automatically rerouted to Claude Opus 4.8. Commerce reviewed that fix before lifting the order.

Is Fable 5 available everywhere now, including on cloud platforms? Yes, on Claude.ai, the Claude Platform, Claude Code, and Claude Cowork immediately, with access on AWS, Google Cloud, and Microsoft Foundry re-enabling on a rolling basis. Mythos 5 was restored on the same schedule.

Could the US government restrict Fable 5 again? Commerce says it retains the authority to reimpose restrictions if circumstances change or if Anthropic doesn’t meet the security commitments tied to the reversal.

Primary sources and further reading

• Redeploying Claude Fable 5 - Anthropic’s official announcement of the July 1 restoration, including the safety classifier details and platform rollout
• Statement on the US government directive to suspend access to Fable 5 and Mythos 5 - Anthropic’s original statement on the June 12 export control order and its response
• Anthropic says Trump admin has lifted export controls on Claude Fable 5 and Mythos 5 - CNBC’s report on the June 30 reversal
• US Lifts Restrictions On Anthropic’s Mythos 5 And Fable 5 AI Models - Forbes coverage with Commerce Secretary Howard Lutnick’s and Susie Wiles’s statements
• Anthropic is bringing back Claude Fable 5 globally after US lifts export control order - VentureBeat on enterprise access paths
• Anthropic Restores Claude Fable 5 After U.S. Lifts Jailbreak-Linked Export Controls - The Hacker News on the security angle
• Claude Fable 5 Restored After U.S. Lifts Export Controls - Bright Defense recap of the timeline
• Anthropic on X: original export control directive statement - Anthropic’s initial public statement on the June 12 suspension