Anthropic re-releases Fable 5 for UK Claude users

UK users of Claude have regained access to Claude Fable 5, Anthropic’s most capable publicly available AI model, after US export controls that cut the model off for nearly three weeks were lifted.

Access was restored globally on Wednesday (1 July 2026) across the Claude Platform, Claude.ai, Claude Code, and Claude Cowork, the AI developer said in an accompanying statement.

Subscribers on Pro, Max, Team, and select Enterprise plans will get Fable 5 included for up to 50% of their weekly usage limits until 7 July, after which the model will be available through usage credits.

Anthropic said it would restore access on AWS, Google Cloud, and Microsoft Foundry as quickly as possible.

Why the model was pulled

Fable 5 launched on 9 June alongside Claude Mythos 5, a version of the same underlying model with fewer safeguards that was restricted to a small group of trusted cybersecurity partners under the company’s Project Glasswing programme.

Three days later, on 13 June, the US government applied export controls to both models, requiring Anthropic to restrict access to foreign nationals inside and outside the United States.

Because the order took effect immediately and the company had no reliable way to verify nationality in real time, it suspended access to both models for all users, including those in the UK.

The directive followed a report in which Amazon researchers found a method of bypassing Fable 5’s safeguards, prompting the model to identify a number of software vulnerabilities and, in one case, produce code demonstrating how a vulnerability could be exploited.

Anthropic said its own testing found that many less capable models, including Claude Opus 4.8, GPT-5.5, and Kimi K2.7, could identify the same vulnerabilities, and that every model it tested could reproduce the single exploit demonstration.

The company characterised the bypass as a borderline case involving routine defensive cybersecurity work rather than an exposure of the more dangerous capabilities reserved for Mythos 5.

New safeguards for the relaunch

Working with the US government, Anthropic trained a new safety classifier that blocks the reported technique in over 99% of cases.

Users will be notified if a request to Fable 5 is blocked, with the request instead routed to Claude Opus 4.8.

The company acknowledged the tighter classifier will flag more benign requests during routine coding and debugging tasks, and said it would continue refining the system to reduce false positives.

Researchers from the US Department of Commerce’s Center for AI Standards and Innovation tested both the prior and new safeguards and agreed they are extraordinarily strong, Anthropic said.

Mythos 5 has been restored only for a set of US organisations following government approval on 26 June, with access for international partners, including any in the UK, still subject to ongoing coordination with the US government.

Industry framework for jailbreaks

Anthropic also announced it is partnering with Amazon, Microsoft, Google, and other Glasswing partners to draft a consensus framework for assessing the severity of AI jailbreaks, techniques used to bypass a model’s safeguards.

The proposed framework would score jailbreaks on four criteria:

• Capability gain: How far beyond existing tools the jailbreak takes an attacker
• Breadth of capability gain: How many distinct offensive tasks the same technique unlocks
• Ease of weaponisation: How much human effort is needed to turn it into an attack
• Discoverability: How easy the technique is to obtain

The company is also launching a HackerOne programme where security researchers can submit potential cyber jailbreaks discovered in Fable 5 for review, and has committed to expanded pre-release government access, rapid information sharing on safeguards, and joint research with US government partners.

Now read: BT bets on AI-ready Verizon venture as Kirkby doubles down on the UK