OpenAI is slowing the launch of its newest AI model.
The company has announced that GPT-5.6 will initially be available only to a small group of trusted partners,
after a request from the U.S. government for additional safety reviews before a wider release.
─────────
The temporary restriction is intended to give authorities time to assess potential national security risks,
including cybersecurity and the misuse of advanced AI capabilities.
─────────
OpenAI says the move is only a short-term measure and expects GPT-5.6 to become
more broadly available in the coming weeks.
However, the company also stressed that government review should not become the
standard process for releasing future AI models.
GPT-5.6 introduces three versions—Sol, Terra, and Luna—designed for different levels
of performance, from advanced reasoning to faster, lower-cost tasks.
─────────
The decision comes as governments around the world increase scrutiny of frontier AI systems,
balancing innovation with concerns over security and public safety.
─────────
One thing is clear.
As AI becomes more powerful, governments are playing an increasingly active role
in deciding how—and when—it reaches the public.
