US Allows Anthropic to Release Mythos 5 to Over 100 Trusted Firms
Source: CNA / Reuters
The US Commerce Department has permitted Anthropic to release its Claude Mythos 5 AI model to over 100 trusted partners after a two-week security review. The authorisation signals a new phase in balancing AI innovation with national security controls.
The United States Commerce Department has authorised Anthropic to release its advanced Claude Mythos 5 AI model to a select group of trusted partners, ending a two-week security review that had blocked the model's broader deployment. Over 100 companies and institutions, including numerous Fortune 500 firms, have been granted initial access under the deal.
The approval, signed by Commerce Secretary Howard Lutnick, follows a June 12 letter that identified specific risks associated with Mythos 5, which Anthropic addressed through a series of security commitments. 'Since the issuance of my June 12 letter, Anthropic has worked with the U.S. government to address risks associated with the Covered Models,' Lutnick wrote, framing the outcome as a successful balance between keeping America's AI lead and protecting national security. The review process has since become a template for how the US government intends to interact with frontier AI developers going forward.
The decision comes amid a broader US crackdown on advanced AI model distribution. OpenAI has similarly deferred the public rollout of its GPT-5.6 at government request, and a customer is now suing the US over lost access to Anthropic's Fable 5 model following export controls. Both OpenAI and Anthropic have confidentially filed for US IPOs, adding a financial dimension to the regulatory pressure.
Why it matters for Singapore: Anthropic has been steadily expanding its presence in Singapore, establishing a local office and actively hiring research and engineering talent in the city-state. As a global hub for AI talent and investment, Singapore is directly affected by US export controls on frontier models. The Mythos authorisation shows that access can be negotiated, but the trend toward government-supervised AI distribution raises questions for Singapore-based developers, enterprises, and researchers who rely on cutting-edge models. The city-state's push to become a trusted AI hub depends partly on maintaining access to frontier AI capabilities while navigating an increasingly fragmented global regulatory landscape.