OpenAI has announced that access to its newest frontier AI model, GPT-5.6 Sol, will initially be limited to a small group of U.S. government-approved users following a request from the Trump administration. The temporary restriction is part of a cybersecurity review aimed at assessing the model's potential risks before a broader public rollout. OpenAI said the measure is intended to support national security while it works with federal agencies on a safe deployment framework.
Under the arrangement, the U.S. government will review and approve early users of the model, which is considered significantly more capable in areas such as cybersecurity and advanced reasoning. Officials are evaluating whether the technology could be misused for cyberattacks, military applications, or other high-risk activities. Broader access is expected only after the review is completed and appropriate safeguards are in place.
OpenAI emphasized that the restricted rollout is a temporary measure and expressed concern that government approval of AI customers should not become a permanent practice. CEO Sam Altman has supported collaboration on AI safety but warned that excessive government involvement in deciding who can access frontier AI models could slow innovation and limit the technology's broader benefits.
The decision reflects a broader shift toward closer government oversight of advanced AI systems in the United States. It follows recent actions involving other leading AI developers and highlights the growing focus on balancing rapid AI innovation with national security, cybersecurity, and public safety as increasingly powerful models enter the market.
