The White House is asking OpenAI to slow roll the release of its new model over safety concerns

2026-06-26

The Trump administration has reportedly pressed OpenAI to limit GPT 5.6 to a vetted, customer-by-customer preview rather than a public launch, mirroring Anthropic's Project Glasswing restrictions on its Claude Mythos cyber model, over concerns that frontier models capable of autonomously finding and exploiting software vulnerabilities pose an outsized risk in the wrong hands.