U.S. Government Sets 30-Day Review Standard for Frontier AI Models

The Trump administration's June AI cybersecurity order, requiring companies to submit frontier models for government review 30 days before public release, has now been applied twice — to OpenAI's GPT-5.6 and Anthropic's Mythos and Fable models. Both labs complied by delaying release and submitting to federal testing. The mechanism mirrors checkpoint infrastructure that typically outlasts its originating circumstances. Whether it becomes permanent depends on how the next model cycle responds to the same request.