Washington, DC – In a move reflecting the US administration’s efforts to balance technological leadership with national security, the US government has approved limited use of the advanced artificial intelligence model “Mythos 5,” developed by Anthropic. This decision comes two weeks after a surprise move to impose strict restrictions on its deployment.
Conditional breakthrough
In a letter to the company, U.S. Commerce Secretary Howard Lutnick stated that Anthropic had made significant progress in addressing the security risks associated with its model. Under this decision, access to its Mythos 5 cybersecurity platform will be made available to a limited number of trusted partners, including cybersecurity experts and critical infrastructure providers.
For his part, Commerce Department spokesman Benoit Cass confirmed that this move is part of Washington’s ongoing efforts to ensure its technological superiority without compromising security standards.
Despite this breakthrough, the fate of the Fable 5 prototype remains uncertain. The department’s directives did not include any relaxation of the restrictions imposed on it. Negotiations between the company and the government are still ongoing to find solutions that guarantee its safe use.
Roots of the crisis and repercussions
This approval comes to defuse the standoff that erupted two weeks ago when the government decided to block foreign access to the two models, fearing they could be circumvented.
Anthropic had warned at the time that applying such stringent standards to all companies could stifle technological innovation. It also emphasized that the “complete resistance” to attempts to bypass security restrictions remains a technological challenge facing the entire sector.
Anthropic wasn’t alone in the eye of this regulatory storm; the Trump administration’s pressure extended to other companies. As a result, OpenAI was forced to restrict the launch of its new GPT-5.6 model.
A wider conflict with the Pentagon
The Mythos 5 crisis is the latest in a series of disputes between Anthropic and US security agencies. The company previously clashed with the Department of Defense (Pentagon) over controls on the use of its models in military applications. This led to the company being designated a “supply chain risk” last March.
Anthropic presents itself as a leader in the responsible development of artificial intelligence. Since the launch of Mythos last April, it has warned of the model’s high vulnerability detection capabilities.
The company is currently working to expedite access for certified partners. Simultaneously, it continues to collaborate with authorities to establish a long-term regulatory framework that prevents a recurrence of such crises and balances technological expansion with sovereign national security requirements.
