OpenAI will delay GPT-5.6 after Trump administration request
TL;DR
OpenAI will reportedly not launch GPT-5.6 broadly right away after a Trump administration request, instead opening only a limited preview for a small group of enterprise customers. Sam Altman reportedly told staff that the US government would approve access customer by customer during the preview, with wider availability possible a few weeks later if the process goes well.
Nauti's Take
The issue is not that safety checks exist. The issue is who controls them, how transparent they are, and whether all labs face the same rules.
A voluntary, improvised preview process may look practical, but it can quickly become a political black box. For OpenAI, this is better than a full halt in the short term, but it sets a risky precedent.
Briefingshow
This is more than a launch delay: if governments pre-approve customer access to frontier models, AI governance moves from broad rules into case-by-case control. That creates uncertainty for companies and potential uneven treatment across competitors. It also shows how quickly national-security logic can turn into industrial policy.
Sources