The U. S.
Department of Commerce approved a broad launch of OpenAI's advanced GPT-5.6 artificial intelligence model on Tuesday, July 7, 2026.
>>> California Judge Dismisses Attempted Murder Charges Against Pasadena Doctor
The decision follows a period of limited testing restricted by government officials.
As reported by Axios, the technology company expects to execute a wide release of the model this week after completing additional testing and final meetings with government representatives.
Background and Previous Restrictions
The policy development follows a decision last month where OpenAI delayed the full public launch of GPT-5.6 at the request of the U.
S. government.
Initial access was limited to 20 vetted organizations approved by the Trump administration.
OpenAI Codex engineering lead Thibaut Sottiaux confirmed on July 6 that the model's highest-compute tier, Sol Ultra, will ship inside the Codex client for trusted API and Codex users.
>>> Sussex Police Appeal for Witnesses After Fatal M23 Crash
The multi-agent system runs on Cerebras chips to process tasks through parallel subagent processes.
It achieved a 91.9% score on Terminal-Bench 2.1 coding workflows, compared to 80.3% scored by Anthropic's Claude Fable 5 on SWE-Bench Pro.
The U. S.
government previously directed competitor Anthropic to suspend access to its Fable 5 and Mythos 5 models under export controls on June 12.
Restrictions were lifted following negotiations and enhanced safety agreements.
Federal scrutiny remains focused on cybersecurity capabilities.
>>> Red Bull F1 Relationship Strained as Performance Clause Triggers
The White House framework established by a June executive order sets voluntary guidelines for advanced frontier models to prevent potential exploitation by hackers.
