Fable 5 Is Back: Anthropic’s Most Powerful AI Returns After U.S. Government Ban

Less than three weeks after disappearing almost overnight, Anthropic’s most powerful artificial intelligence model is making its comeback.

The company announced that Claude Fable 5 will begin rolling out again after the U.S. Department of Commerce lifted the export controls that forced Anthropic to suspend access to both Fable 5 and its even more powerful sibling, Mythos 5. The decision ends one of the most dramatic episodes the AI industry has witnessed this year and raises fresh questions about how governments will regulate frontier AI models going forward.

Anthropic confirmed the news in a statement, thanking customers for their patience and saying it had been working closely with U.S. officials to restore access. The rollout begins immediately, with availability expanding across Anthropic’s platform and cloud partners in the coming days. 

When Anthropic launched Claude Fable 5 in June, it was billed as the company’s most capable public model yet, built on its new Mythos-class architecture.

Three days later, everything changed.

The U.S. government issued an export-control directive requiring Anthropic to suspend access after concerns emerged that sophisticated jailbreak techniques could potentially expose advanced cybersecurity capabilities to unauthorized users. Because Anthropic could not immediately distinguish eligible users from restricted ones worldwide, the company temporarily disabled both Fable 5 and Mythos 5 for everyone.

The shutdown sent shockwaves through the AI community.

Developers who had just begun experimenting with Fable 5 suddenly found the model unavailable, while enterprises that had planned deployments were forced back onto older Claude models.

Anthropic says it has now implemented additional safeguards that satisfied the U.S. government.

Among the changes are new systems designed to detect jailbreak attempts, expanded monitoring for suspicious activity, closer information sharing with U.S. authorities and participation in industry-wide security initiatives alongside major cloud providers. The company also says it has introduced new internal monitoring teams and expanded its vulnerability reporting programme.

Anthropic has also acknowledged an uncomfortable reality.

No frontier AI model can be made completely immune to jailbreaks.

Instead, the company says its goal is to make successful attacks significantly more difficult while responding rapidly whenever new techniques are discovered.

The return of Fable 5 is about much more than one AI model.

It highlights how governments are becoming increasingly involved in determining when the world’s most powerful AI systems can be released.

Just a few years ago, companies largely decided for themselves when new models were ready for public use.

Today, the situation is changing.

As AI systems become more capable in areas such as cybersecurity, software engineering and scientific research, regulators are demanding greater visibility into how these models are tested, deployed and monitored.

Anthropic’s agreement to work more closely with U.S. authorities may become a blueprint for how other frontier AI developers—including OpenAI, Google DeepMind and xAI approach future model releases.

Fable 5 isn’t simply another chatbot.

It represents Anthropic’s next generation of frontier AI, offering major improvements in reasoning, software development and complex problem-solving compared with previous Claude models.

Its release was intended to showcase the company’s ability to compete with OpenAI’s GPT family, Google’s Gemini models and xAI’s Grok. Instead, it quickly became the centre of an international debate about AI safety and national security.

Now, with restrictions lifted, developers can once again explore what many consider one of the most capable AI systems currently available.

The brief disappearance of Fable 5 may ultimately be remembered as the moment governments demonstrated they were willing to intervene directly in the deployment of frontier AI.

For AI companies, the message is becoming increasingly clear.

Building the world’s smartest models is no longer enough.

Winning the trust of regulators may soon become just as important.

For developers, enterprises and researchers, Fable 5’s return is welcome news.

For the AI industry as a whole, however, it is another reminder that the race to build increasingly powerful artificial intelligence is now being shaped as much by policy as by technology.