OpenAI launches GPT-5.6 after U.S. delay
OpenAI launched GPT-5.6 on Thursday after postponing the public release at the request of U.S. officials; it also unveiled lower-cost Terra and Luna models.
OpenAI launched GPT-5.6 on Thursday after postponing the public release last month at the request of U.S. government officials. The company introduced GPT-5.6 Sol as its most capable model and announced lower-cost alternatives called Terra and Luna.
Before the public launch, OpenAI limited access to GPT-5.6 to a small group of vetted partners and shared information about those partnerships with U.S. agencies while carrying out additional tests. Federal authorities had raised concerns that increasingly capable systems could be repurposed by foreign governments or military organisations.
OpenAI posted on X that GPT-5.6 Sol would be the flagship and that Terra and Luna were intended to widen access through lower pricing. The company said it took extra steps to address safety and security questions before the wider release.
The U.S. has a voluntary framework that allows developers to submit frontier models to federal reviewers for up to 30 days before releasing them to trusted partners. The process is intended to give officials time to assess possible risks and coordinate safeguards when needed.
Other AI firms have adjusted access to advanced systems amid regulatory pressure. One company temporarily disabled its most advanced models after a U.S. export control order in June; some restrictions were later eased after the company introduced additional safeguards. Certain specialised systems, such as models built for cybersecurity tasks, remain limited to a set of trusted U.S. organisations.
The launch comes as the United States and China compete over advanced AI capabilities. Chinese authorities have met with domestic technology firms to discuss possible limits on overseas access to the country’s most capable models. Governments on both sides have taken steps to review or restrict distribution of frontier systems for national security and cybersecurity reasons.
OpenAI’s rollout, with early access for vetted partners followed by a public release, follows a pattern used by other developers to test advanced systems and coordinate with regulators before broader deployment.








