Claude Fable 5 is returning after US export controls are lifted

Explore with AI:

Anthropic says Claude Fable 5 will return globally starting Wednesday, July 1, after the US government lifted export controls on both Fable 5 and Mythos 5.

Fable 5 will be available again on Claude Platform, Claude.ai, Claude Code, and Claude Cowork. Anthropic also says it will restore access through AWS, Google Cloud, and Microsoft Foundry as quickly as possible, but it has not given a fixed timeline for those cloud platforms yet.

This ends a strange three-week disruption for one of Anthropic’s most important public models. The company had launched Claude Fable 5 on June 9 as its first broadly available Mythos-class model, then pulled it offline on June 12 after a US export control directive covered access by foreign nationals.

Fable 5 access comes back with limits

For Claude subscribers, the return is not exactly the same as a normal model rollout. Anthropic says Fable 5 will be included for up to 50% of weekly usage limits on Pro, Max, Team, and select Enterprise plans through July 7. After that, users will need usage credits.

Mythos 5 is still being handled more cautiously. Anthropic says access has been restored for a set of US organizations after government approval on June 26, and the company is still working to expand access to more domestic and international partners in its Glasswing program.

That distinction matters. Fable 5 is the safer general-use version. Mythos 5 uses the same underlying model, but Anthropic has positioned it for trusted cybersecurity partners rather than everyday users.

The jailbreak fight is not over

The export-control order followed government concern over a reported method for bypassing Fable 5’s safeguards. Anthropic says Amazon researchers had found a way to prompt Fable 5 into identifying software vulnerabilities, including one case where it produced exploit-demonstration code.

Anthropic now says it has trained an improved safety classifier that targets the specific behavior described in that report. If Fable 5 blocks a request, the user will be notified and the request will be routed to Opus 4.8 instead.

The company says the new classifier blocks the reported technique in more than 99% of cases. That does not mean Fable 5 is jailbreak-proof. Anthropic is also proposing a shared industry framework, with Amazon, Microsoft, Google, and other Glasswing partners, to score AI jailbreaks by severity and decide how quickly companies should respond.

Why Apple users should care

Claude is no longer just another chatbot brand. Anthropic has been pushing it into more work surfaces, including Claude Tag for Slack, while Apple is reportedly exploring broader third-party AI model choices for Apple Intelligence.

That makes access to flagship models like Fable 5 worth watching even if you do not use Claude every day. As AI assistants move deeper into phones, Macs, work apps, and coding tools, government rules around who can access the most capable models could start shaping what users actually get to use.

Explore with AI: