The United States has lifted the requirement for Anthropic to obtain a license before exporting its Mythos and Fable models overseas. This requirement effectively cut off public access to what is widely considered the most advanced AI model ever released.
AI Lab announced that it will begin restoring access to its models on Wednesday, July 1st.
On June 12, the U.S. government added these products to its list of export-restricted technologies. This means that those products will be available to foreigners without special approval. Following this rule proved impractical at scale, and Anthropic was forced to completely end public access to the model.
After weeks of discussions, Secretary of Commerce Howard Lutnick has now said that Anthropic has “agreed to proactively detect and address security risks associated with the models, to work diligently with the U.S. government on Mythos, Fable, and future model protocols and standards and releases, and to notify the U.S. government of any malicious activity.”
Anthropic had already publicly committed to voluntarily undertake many of these efforts months before export restrictions existed. That’s part of the reason cybersecurity experts were skeptical of the limit in the first place. To them, the ban looked less like a security measure and more like a way for the Trump administration to punish Anthropic for its management’s public criticism of how the technology is used by the government and the president’s political opponents.
Mythos was originally made available to a select group of organizations in April to allay concerns about the software’s ability to identify and exploit vulnerabilities, but a version called Fable with additional security guardrails was made publicly available in June.
However, as Asian AI companies began releasing their own AI models approaching Mythos-level capabilities, including Fugu and Tulonfeng, the US government was under pressure to ease restrictions on Anthropic to allow US AI to compete globally.
Last week, Lutnick authorized the release of Mythos to select customers approved by the White House. OpenAI’s latest model was also not released to the public, but to a group of organizations approved by the Trump campaign.
The Trump administration’s erratic approach to AI policymaking has left companies across the industry with little clarity on how to decide on future model releases. An executive order issued in June indicating the company’s intention to review models before release was criticized by influential analysts like Dean W. Ball, who recently became director of policy at OpenAI.
If you buy through links in our articles, we may earn a small commission. This does not affect editorial independence.
