Anthropic’s Claude Mythos 5 AI Model Receives Government Approval for Limited Access

On Friday, Anthropic Inc. announced that the U.S. government has permitted it to provide access to its powerful Claude Mythos 5 AI model to certain “trusted” organizations, reversing a two-week suspension due to national security concerns.

Reports suggest that over 100 companies and institutions, including several Fortune 500 firms, will now have access to Mythos 5, according to unnamed sources familiar with the situation.

The decision to restrict AI models stems from fears of misuse by military intelligence in nations like China and Russia, prompting the Trump administration to adopt strict oversight over the releases from both Anthropic and competing firm OpenAI.

OpenAI also announced a delay in the full release of its GPT-5.6 model, complying with government requests to limit access to a small group of vetted partners, the specifics of which have been shared with authorities.

Following a government export control order issued on June 12, Anthropic had temporarily disabled its recent AI models, including Mythos 5 and Fable 5, for all users.

In its statement, Anthropic mentioned that it received notification from the government allowing the deployment of Mythos 5 to several U.S. organizations responsible for critical infrastructure protection.

The company is in the process of restoring access to these entities and plans to collaborate with the government to broaden availability for Mythos 5 and potentially make Fable 5 accessible again.

Criticism has emerged regarding the government’s criteria for selecting companies allowed to access Mythos. John Coleman, legislative counsel for the Foundation for Individual Rights and Expression, expressed concerns about the lack of transparency, questioning the fairness of the selection process.

In a post on X, OpenAI CEO Sam Altman shared similar worries, stating that while extensive safety testing is a good idea, he disapproves of governmental control over access to advanced models.

Experts point out that, if misused, the Mythos model could significantly enhance cyberattack capabilities, particularly in sectors like banking that depend on complex, aging technologies.

A letter from U.S. Secretary of Commerce Howard Lutnick highlighted that Anthropic had made notable progress in addressing government concerns regarding the risks linked to its targeted models.

Details about specific safety measures taken remain unclear. Earlier this month, Anthropic acknowledged that the government had concerns about potential circumvention of safeguards meant to prevent Fable 5 from identifying software vulnerabilities.

Lutnick’s letter indicated that export licenses for Mythos 5 will no longer be necessary for trusted companies and non-U.S. citizen employees, although restrictions will persist for businesses not included on the approved list.

Many of these approved firms are participants in Anthropic’s Project Glasswing, which encompasses about 100 major tech companies and institutions, according to sources.

Additionally, the government may soon grant Anthropic the go-ahead to release “Fable,” although a definitive timeline has not been established.

Fable 5 and Mythos share the same core AI model; however, Fable 5 is intended for broader public access, whereas Mythos has certain safeguards relaxed.

Both Anthropic and OpenAI have plans to pursue public offerings in the future.

Anthropic’s interactions with the U.S. government have been particularly complex. The company has previously declined to let the military utilize its AI models for domestic surveillance or fully autonomous weapons, which resulted in the government placing it on a national security blacklist.

The restrictions imposed on Anthropic and OpenAI followed an executive order signed by President Trump, aimed at establishing a voluntary framework for AI developers to provide “covered frontier models” to the U.S. government for up to 30 days prior to releasing them to trusted partners.

Kate Koren, an analyst at the Center for Strategic and International Studies with prior experience at the Commerce Department, noted that the administration’s latest order serves as a temporary solution but raises lingering questions about how companies can launch their latest models more broadly.

She remarked that as U.S. companies face delays in releasing new models, the risk that China could catch up increases.