Commerce Department Clears OpenAI GPT-5.6 for Full Public Release After Staggered Review

OpenAI’s Sol flagship plus Terra and Luna variants were initially restricted in June to government-approved entities under the administration’s AI executive order. The company had announced a staggered rollout but stated it was not preferred policy. After OpenAI engineers met CAISI reviewers and additional safety tests were completed, the models received clearance for Thursday launch with expanded global preview access.

June export controls on Anthropic’s Mythos and Fable models barred foreign nationals and forced temporary market withdrawal. Fable access was restored last week after guardrail upgrades that reduced reported performance. These actions establish model-by-model negotiation between frontier labs and Commerce rather than uniform standards still under development.

The United States gains continued oversight of leading closed models to address misuse and security risks but incurs coordination delays that slow domestic deployment. Open-weight Chinese systems face fewer release frictions and are expanding market share. Beijing is simultaneously drafting its own access restrictions, creating symmetric controls on both sides.

Standardized release frameworks called for in the executive order remain unfinished. Developers can now access full GPT-5.6 capabilities while regulators refine the precedent for future frontier systems.