The United States government has lifted export restrictions on Anthropic’s most advanced AI tools, only a few weeks after it had ordered the company to limit access due to national‑security worries, Anthropic reported.
Anthropic announced via a social‑media post that it will start restoring access to Claude Fable 5 and Mythos 5 on Wednesday, after receiving notice that the U.S. Department of Commerce has lifted the export controls on those models.
These represent the company’s most advanced AI offerings, which were abruptly suspended on June 12 after authorities warned they could be leveraged by hackers to find and exploit system vulnerabilities.
The BBC has reached out to the Department of Commerce for a statement.
Mythos and Fable are two of Anthropic’s AI models constructed on its Claude platform, which competes with OpenAI’s ChatGPT and Google’s Gemini.
Fable 5 is a consumer‑focused version of the AI model, capable of deep reasoning and able to perform complex tasks independently.
Mythos 5 is a variant aimed at business users and cybersecurity professionals, reportedly capable of spotting vulnerabilities in code and exploiting them.
The company previously noted that U.S. officials had not identified specific issues with its technology, even while ordering the global suspension of both platforms.
“Our understanding is that the government believes it has uncovered a way to bypass, or “jailbreak,” Fable 5,” the company said at the time, referring to a technique that evades software safety controls to unlock restricted features.
“However, we contend that the discovery of a limited potential jailbreak does not justify withdrawing a commercial model used by hundreds of millions of people.”
