OpenAI has announced a phased, limited rollout for its upcoming GPT-5.6 model, following requests from the US federal government. The company delayed the full public launch at the request of authorities who shared details with vetted partners only.
This decision highlights growing concerns in Washington over national security risks posed by powerful AI systems. Policymakers are urging companies to implement safeguards around these technologies.
The rollout comes as OpenAI complies with security requests, aligning with the Trump administration’s framework for frontier AI safety. Early access to frontier models aims to identify potential threats from cyberattacks to military misuse before widespread deployment.
OpenAI stated in a blog post that the limited release is a temporary step while working with Washington on a broader framework for future launches. The company presented its plans and model capabilities to the government prior to launch, according to OpenAI.
The executive order signed by President Donald Trump earlier this month established a voluntary framework for AI developers to offer “covered frontier models” to the US government for up to 30 days before releasing them to trusted partners.
OpenAI expressed concern that such a process could restrict access to advanced AI tools for users including developers, businesses, cybersecurity professionals and international partners who could benefit from them. The new lineup includes GPT-5.6 Sol, the company’s most advanced model yet, alongside Terra and Luna.