The decision highlights growing concerns in Washington about the national security risks posed by powerful AI systems.
issued Saturday, June 27, 2026 · 02:58 PM
[BENGALURU] OpenAI announced on Friday (June 26) that it will postpone the full rollout of GPT-5.6 at the request of the US government and will restrict initial access to the AI model to a small group of vetted partners whose details have been shared with authorities.
The decision highlights growing concerns in Washington about the national security risks posed by powerful AI systems, and policymakers are pressuring companies to put guardrails in place.
By securing early access to the frontier model, U.S. officials aim to identify threats ranging from cyberattacks to military exploitation before the tools are widely deployed.
OpenAI said in a blog post that the limited rollout is a temporary measure to work with Washington on a broader framework for future rollouts. The ChatGPT manufacturer added that it had presented its plans and model features to the government before the launch.
CEO Sam Altman told X that extensive safety testing “isn’t a bad idea. We just don’t like the idea of the government choosing our customers.”
President Donald Trump signed an executive order earlier this month establishing a voluntary framework for AI developers to provide “targeted frontier models” to the U.S. government for up to 30 days before making them available to trusted partners.
OpenAI said: “We are taking this short-term step because we believe this is the strongest path to achieving broader availability in the coming weeks. Meanwhile, we are working with the government to develop a cyber executive order framework and a repeatable process for future model releases.”
The company said it would continue rigorous testing and close coordination with partners in preparation for widespread release, but cautioned that this level of government access and oversight should not become a permanent norm. The partner’s name was not disclosed.
However, OpenAI expressed concern that such a process would limit access for users who could benefit from advanced AI tools, including developers, enterprises, cybersecurity experts, and international partners.
Related items
The centerpiece of the new lineup is OpenAI’s most advanced model to date, the GPT-5.6 Sol, alongside the mid-tier Terra and budget Luna.
Earlier this month, the U.S. government ordered Anthropic to cut off access to its Frontier AI models for foreigners, citing national security concerns. But on Friday, the government partially rescinded and granted the AI developer permission to release its advanced Claude Mythos 5 model to a select group of trusted American organizations.
OpenAI and Anthropic have both secretly filed for initial public offerings in the United States. new york times We reported on Thursday that OpenAI is considering delaying its public debut until next year. Reuters
