OpenAI has postponed the full public release of its next-generation GPT-5.6 artificial intelligence models at the request of the U.S. government, highlighting growing federal scrutiny over the national security implications of increasingly powerful AI systems.
Instead of a broad rollout, the ChatGPT developer said GPT-5.6 will initially be made available only to a small group of vetted partners whose identities have been shared with U.S. authorities, pending the development of a broader government framework for future frontier AI releases.
The move follows increasing efforts by Washington to strengthen oversight of advanced AI technologies amid concerns they could be exploited for cyberattacks, military applications or other national security threats.
In a blog post, OpenAI said the limited release is a temporary measure designed to support collaboration with the U.S. government while allowing additional safety testing before wider public availability.
“We are taking this short-term step because we believe it is the strongest path to broader availability in the coming weeks, while we work with the administration to develop the cyber executive order framework and a repeatable process for future model releases,” the company said.
However, OpenAI also cautioned against making government-controlled access a permanent feature of AI development, warning that it could unnecessarily restrict access for developers, businesses, cybersecurity professionals and international users.
OpenAI Chief Executive Officer Sam Altman echoed that concern in a post on X, stating: “Extensive safety testing is not a bad idea. I just don’t like the idea of the government picking the customers.”
The delay comes after U.S. President Donald Trump signed an executive order earlier this month establishing a voluntary framework under which AI developers can provide “covered frontier models” to the federal government for up to 30 days before wider release to trusted partners.
At the centre of OpenAI’s latest product lineup is GPT-5.6 Sol, described as the company’s most advanced AI model to date, alongside the mid-tier Terra and lower-cost Luna models.
The development reflects a broader shift in U.S. AI policy. Earlier this month, the government ordered rival AI company Anthropic to suspend access to its frontier models for foreign nationals over national security concerns.
That restriction now appears to be easing. Anthropic announced that it has been authorised to release its Claude Mythos 5 model to selected trusted U.S. organisations, while reports indicate the Trump administration is also preparing to allow public access to its Fable 5 model in the coming days.
The evolving regulatory landscape underscores the increasing tension between accelerating AI innovation and ensuring advanced models do not pose risks to national security, with Washington taking a far more active role in determining how frontier AI technologies are deployed.
