June 26 (Reuters) - OpenAI said on Friday it would delay the full public launch of its GPT-5.6 model at the request of the U.S. government, limiting initial access to the AI model to a small group of vetted partners whose details were shared with authorities.

This decision underscores growing concern in Washington about the risks that powerful artificial intelligence systems pose to national security, as policymakers pressure companies to put controls in place to regulate them.

By ensuring early access to advanced models, U.S. officials aim to identify threats ranging from cyberattacks to military misuse before deploying these tools widely.

OpenAI said in a blog post that the limited release is a temporary step, as it works with Washington to develop a broader framework for future releases. The ChatGPT developer added that it had presented its plans and model capabilities to the government prior to the launch.

OpenAI said, "We are taking this short-term step because we believe it is the strongest path toward providing the models more broadly in the coming weeks, while we work with the U.S. administration to develop its cybersecurity executive framework and a repeatable process for future model releases."

The company added that it will continue to conduct rigorous testing and coordinate closely with its partners as it prepares for a wider release, but cautioned that this level of government access and oversight should not become the norm. The company did not disclose the names of its partners.