Trump administration forces OpenAI to gate GPT 5.6
openai's gpt 5.6 release will be restricted to a select group of partners after the trump administration demanded safety review. this marks a shift in federal oversight, giving the government approval power over customer access.
openai ceo sam altman told staff that the trump administration is blocking the public release of gpt 5.6, instead requiring a limited rollout where the government reviews and approves access customer by customer. agencies including the office of the national cyber director reportedly intervened over concerns that the model could accelerate cyberattacks, mirroring anthropic's controversial strategy of restricting its claude mythos model. this intervention follows an executive order requiring voluntary red-teaming, but represents a significant escalation by imposing specific deployment constraints. by mandating a staggered release, the government establishes a new precedent where national security concerns directly delay and dictate the terms of ai model distribution.