Anthropic wants to make a deal. The AI company behind the Claude family of models proposed working directly with the Trump administration on June 18, 2026, to resolve cybersecurity concerns and end a US ban on its most advanced AI products, Mythos 5 and Fable 5.
The proposal comes just five days after Anthropic was forced to suspend access to both models following an export control directive from the administration. That suspension didn’t just affect foreign users. It cut off access for everyone, including US employees, as the company scrambled to comply with the government’s order.
How we got here
On June 12-13, the Trump administration issued an export control directive prohibiting foreign nationals from accessing both Mythos 5 and Fable 5. The stated reason: cybersecurity concerns tied to the models’ advanced capabilities.
By June 13, Anthropic had pulled the plug entirely, suspending all access to both models.
Two days later, on June 15, senior Anthropic executives began meeting with Trump administration officials. Commerce Secretary Howard Lutnick has been a key figure in those discussions.
Then came June 18, when Anthropic formally proposed a closer collaborative framework. Anthropic is also actively disputing the severity of the cybersecurity risks the administration cited when it imposed the ban.
What Mythos and Fable actually are
Mythos 5 launched earlier in 2026 and represents Anthropic’s most capable model to date, with particularly notable performance in cybersecurity applications, including identifying and mitigating vulnerabilities.
Fable 5 offers a more user-friendly experience with additional safeguards designed to prevent the propagation of risky outputs, making it more accessible to a wider range of users.
Cybersecurity experts have argued that restricting access to top-tier American AI models does not weaken adversaries but reduces the competitive edge of American companies.