The United States government has officially granted Anthropic permission to provide its advanced artificial intelligence model, Claude Mythos 5, to a restricted group of trusted cybersecurity firms in the US. The company confirmed the development on Friday, marking a shift in the government’s stance regarding the powerful AI technology.
In a formal statement released on Friday, Anthropic noted that the government had notified them that Mythos 5, which serves as the company’s most robust cybersecurity tool, could be redeployed to a set of American organizations tasked with operating and defending critical infrastructure. This update comes exactly two weeks after the Trump administration initially moved to block access to the software.
The company is currently working to restore access for these designated organizations as quickly as possible. A spokesperson for Anthropic highlighted that the firm remains in active discussions with the government to eventually expand access to Mythos 5 and to make the Fable 5 model available to the general public once again. The company’s relationship with the Trump administration has been notably strained for several months.
Previously, on June 12, Anthropic was forced to pull both Mythos 5 and Fable 5 from all users following a government order restricting their release. Officials at the time cited concerns that such sophisticated systems could potentially be misused by military intelligence entities in countries like Russia or China, prompting a rigorous approach to oversight for both Anthropic and its competitor, OpenAI.
While the Trump administration faced accusations of government overreach regarding these restrictive measures, it has yet to disclose the specific criteria for vetting the organizations or which companies will ultimately receive access to the model. OpenAI CEO Sam Altman expressed reservations about the government’s role in selecting clients, stating that while thorough safety testing is beneficial, he does not support the government choosing who gets to use top-tier models.
According to an anonymous source familiar with the new directive speaking to Reuters, more than 100 institutions and companies—including numerous Fortune 500 businesses—are expected to gain access to Mythos 5 under the new guidelines.
