Key facts
- US approves limited — not full — access to Anthropic's Mythos AI model
- Model was previously blocked over national security concerns
- Decision follows a formal security review process
- Restricted rollout, not open public release
The United States has approved limited access to Anthropic's Mythos AI model, reversing an earlier block that was imposed on national security grounds, multiple Indian and international outlets reported on Thursday. The rollout is restricted in scope, meaning the model will not be freely available but will instead be accessible under controlled conditions following a formal security review.
The move reflects a broader pattern in Washington's approach to frontier AI development, where models deemed sufficiently powerful are subject to government scrutiny before release. Anthropic, one of the leading AI safety-focused research companies, had previously seen access to the model curtailed while authorities assessed potential risks. The limited clearance signals that the review found the model could be deployed under appropriate safeguards.
The development has significant implications for the global AI ecosystem. By imposing access controls rather than outright bans, the US government is carving out a framework in which it acts as a gatekeeper for the most advanced AI systems — a stance that could influence how other nations, including India, seek to access or develop similar technologies.
For now, the approval is partial. Anthropic and users seeking access to Mythos will operate within the boundaries set by the security review's findings, with broader release remaining contingent on further clearances or evolving policy decisions.