WASHINGTON ( Web Desk) — The United States has partially lifted restrictions on Anthropic’s advanced Claude Mythos 5 artificial intelligence model, allowing more than 100 trusted American organisations to regain access after a temporary suspension imposed over national security concerns.
Anthropic announced on Friday that the US government had authorised the redeployment of Mythos 5 to selected organisations responsible for operating and defending critical infrastructure. The company said it is working with federal authorities to expand access further and eventually restore public availability of its Fable 5 model.
The move follows a June 12 export control order that halted access to Anthropic’s most advanced AI models amid concerns they could be exploited by hostile governments or cybercriminals. Officials have warned that highly capable AI systems could accelerate sophisticated cyberattacks, particularly against sensitive industries such as banking and critical infrastructure.
According to a source familiar with the decision, the approved organisations include many Fortune 500 companies, with several participating in Anthropic’s Project Glasswing initiative, a programme involving leading technology firms and research institutions.
The announcement came shortly after OpenAI revealed it had postponed the full public launch of GPT-5.6 at the request of the US government. The company said the model would initially be available only to a limited group of vetted partners while additional safety reviews are conducted.
The government’s role in deciding which organisations receive access to advanced AI models has drawn criticism from free speech advocates and technology leaders.
John Coleman, legislative counsel for the Foundation for Individual Rights and Expression, questioned the lack of transparency surrounding the approval process, arguing that it gives the government excessive influence over access to emerging technologies.
OpenAI Chief Executive Sam Altman also expressed concern, saying that while rigorous safety testing is reasonable, the government should not determine which customers can use frontier AI models.
US Commerce Secretary Howard Lutnick said the government had made significant progress with Anthropic in addressing security risks associated with the company’s most powerful AI systems. Under the revised policy, approved organisations and their non-US employees will no longer require export licences to use Mythos 5, although restrictions remain in place for companies outside the approved list.
Anthropic previously said the government believed certain safeguards protecting Fable 5 from misuse could potentially be bypassed through advanced “jailbreaking” techniques. It remains unclear what additional security measures have since been implemented.
Both Mythos 5 and Fable 5 are built on the same underlying AI technology. However, Mythos 5 is designed for specialised cybersecurity applications with fewer operational restrictions, while Fable 5 is intended for broader public use.
The latest policy reflects the Trump administration’s broader effort to tighten oversight of powerful AI systems. Earlier this month, President Donald Trump signed an executive order creating a voluntary framework allowing developers to provide frontier AI models to the US government for review before wider deployment.
Analysts say the temporary restrictions may help address immediate security concerns but warn that prolonged limits on releasing advanced AI could weaken the competitiveness of US technology companies as rivals, particularly China, continue to invest heavily in artificial intelligence.
