AI & Models
Anthropic regains access to Mythos 5 for select US organizations
The Trump administration is partially lifting its ban on Anthropic’s Mythos 5, allowing more than 100 US organizations to redeploy the cybersecurity-oriented AI model.
The Trump administration is softening its stance on the ban of Anthropic’s cybersecurity-oriented models. Under a new directive, the artificial intelligence developer is permitted to make its Mythos 5 model available to more than 100 specific US government agencies and companies. The decision, which was reported by both Semafor and Reuters, marks a partial easing of a regulatory ban that had previously forced the company to pull its specialized security models from the market.
The new authorization specifically extends to non-American employees working at the authorized organizations. This permission addresses a key restriction of the original ban, which had barred non-American personnel—including Anthropic’s own non-American staff—from accessing the technology. However, the administration did not address the release of Fable 5 in the directive, meaning that model remains under the original restrictions.
The regulatory rollback follows safety concerns that arose when both models were pulled after their guardrails were allegedly bypassed easily by security researchers. Following the pull of the models, Anthropic began working closely with the US government on June 12 to restore access to both Claude Mythos 5 and Fable 5.
On Friday, Commerce Secretary Howard Lutnick—the US official overseeing trade and business regulation—formally authorized the partial redeployment in a letter written to Tom Brown, Anthropic’s chief compute officer. In the letter, Lutnick wrote: “I have determined that appropriate safeguards are in place to permit certain trusted partners to access the Claude Mythos 5 Model.”
Anthropic acknowledged the decision, noting that the US government notified the company that Mythos 5, which it describes as its strongest cybersecurity model, can now be redeployed to a set of US organizations that operate and defend critical infrastructure. The company stated that it is restoring access for these specific organizations quickly, and it continues to work with the government to expand access to Mythos 5 and make Fable 5 available for general use again.
Why it matters
This partial reversal signals a nuanced regulatory approach to AI safety, balancing the need for critical infrastructure protection with the risks of powerful cybersecurity models.