Why Fable 5 Disappeared From Claude in the First Place

The trouble started with a government order, not a technical failure. A directive arrived requiring Anthropic to block foreign nationals from using Claude Fable 5 and its more capable sibling, Mythos 5. The catch was timing: the rule took effect immediately, and Anthropic had no way to check a user's nationality in real time. Rather than risk a violation, the company pulled both models for everyone, everywhere, all at once.

The Vulnerability That Started It All

The shutdown traces back to a report from Amazon researchers. They found a way to prompt Fable 5 into naming software vulnerabilities that its safeguards were supposed to keep locked down. In at least one case, the model didn't stop at identifying the flaw — it generated code showing how that flaw could actually be exploited. That was enough to trigger a full suspension while regulators and Anthropic worked out what to do next.

Full Access Comes Back Starting July 1

Fable 5 is available globally again starting July 1, and this time it's rolling out across the Claude Platform, Claude.ai, Claude Code, and Claude Cowork all at once. Anthropic confirmed the news itself, posting that it had "received notice that the Department of Commerce has lifted export controls" and would begin restoring access the next day.

Worth noting: Anthropic says the jailbreak technique that caused all this wasn't unique to Fable 5 in the end. The company later confirmed that other, less capable models — including Opus 4.8 and rival models from other labs — could be pushed into producing similar results. That detail matters, because it suggests the bypass had more to do with a general weakness in how models handle certain prompts than with anything specific to Fable 5's added capabilities.

Usage Limits During the Rollout

If you're on a Pro, Max, Team, or select Enterprise plan, there's a temporary catch. Fable 5 usage will count toward up to 50 percent of your weekly usage limit through July 7. After that date, Fable 5 usage shifts over to draw from usage credits instead.

A New Safety Classifier Is Doing the Heavy Lifting

Getting Fable 5 back online wasn't just a matter of flipping a switch. Anthropic built a new safety classifier, developed in coordination with the government, specifically to target and block the technique that caused the shutdown. The company says this classifier catches the flagged method in more than 99 percent of cases.

Here's how it works in practice: if a request to Fable 5 gets blocked, the user is notified, and the query gets automatically rerouted to Opus 4.8 instead. Anthropic has also signaled that the new classifiers are tuned to block a broader range of cybersecurity-related tasks going forward, which means some routine work — everyday coding included — could occasionally get caught up in the net in the near term.

Where Mythos 5 Stands Right Now

Mythos 5, the more capable of the two models, isn't back to full strength yet. Its access was restored on June 26, but only for a limited set of approved U.S. organizations following a separate government review. Anthropic says it's continuing to work with the government to widen that access to more domestic and international partners over time.

What Comes Next

This isn't just a one-off fix. Anthropic is now working alongside Amazon, Microsoft, and Google to build a shared industry standard for grading how severe an AI jailbreak actually is. That kind of framework could end up shaping how the entire industry responds the next time a safeguard gets bypassed.