Claude Fable 5 Is Back: Export Controls Lifted, Global Access Returns

Claude Fable 5 is back. The US Department of Commerce lifted the export controls that forced Anthropic to pull Fable 5 and Mythos 5 offline on June 12, and Anthropic began restoring global access on July 1.

If you had workflows built on either model, here's the full timeline and what actually changed.

The Three-Week Timeline

June 9: Anthropic launched Claude Fable 5 for general users and Claude Mythos 5 for a limited group of trusted Project Glasswing cybersecurity partners. Both share the same underlying model, but Fable shipped with stronger safety guardrails.

June 12: I covered this as it happened — the government ordered Anthropic to suspend access for any foreign national, inside or outside the US, including Anthropic's own non-citizen staff. Anthropic had no real-time way to verify every user's nationality, so it shut both models down for everyone.

The trigger was a jailbreak. Amazon researchers found a prompt that got Fable 5 to flag software flaws and, in one case, write code demonstrating how a flaw could be exploited. Anthropic downplayed the severity, saying it involved a small number of previously known, minor vulnerabilities — not a real breach.

June 26: The government partially reversed course, allowing Mythos 5 access to return for roughly 100 trusted US companies and federal agencies working on critical infrastructure defense.

June 30: Commerce Secretary Howard Lutnick sent Anthropic a letter withdrawing the export-control license requirement entirely for both models.

July 1: Fable 5 returned globally across Claude.ai, the Claude Platform, Claude Code, and Claude Cowork. Mythos 5 stays restricted to the roughly 100 approved US organizations, with Anthropic still negotiating wider access.

What Anthropic Actually Fixed

Anthropic says it trained a new safety classifier specifically to catch the bypass technique Amazon found, after concluding the jailbreak didn't expose true Mythos-level offensive capability. The updated classifier reportedly blocks the reported technique in more than 99% of cases. When it fires, the user gets a notification and the request gets rerouted to Claude Opus 4.8 instead.

Anthropic also opened a HackerOne program specifically for researchers to report new Fable 5 jailbreaks, and stood up a 24/7 monitoring team for the model.

What Changes for Anthropic Going Forward

The bigger shift is process, not just this one model. Anthropic says it will now give designated government partners early access to frontier models and their safeguards for independent testing before public release, and will rapidly notify agencies about significant jailbreaks and new safeguards going forward.

That lines up with a June 2 executive order pushing federal agencies toward a formal process for benchmarking and assessing new frontier models before wide release, with a 30-day review window. Anthropic isn't alone here either — OpenAI previewed GPT-5.6 to a small government-approved group rather than the public, citing the same dual-use concern: a model good enough to help defenders patch bugs is also good enough to help attackers find them.

What This Means If You Build on Fable 5

Access is back across Claude.ai, the Claude Platform, Claude Code, and Claude Cowork as of July 1. Pro, Max, Team, and select Enterprise plans get Fable 5 included at 50% of normal usage limits through July 7 — after that, it shifts to usage-based credits. Standard Enterprise seats have no included Fable 5 allowance and need credits enabled to use it at all.

Cloud availability is rolling out in phases — Anthropic says it's re-enabling Fable 5 on AWS, Google Cloud, and Microsoft Foundry "as quickly as possible," but it wasn't live everywhere as of this writing. If your pipeline depends on one of those platforms specifically, check before you assume it's back.

Sources: Anthropic, The Hacker News