GPT-5.6: OpenAI gets the green light from Donald Trump; the new ChatGPT arrives this Thursday.

Last month, OpenAI discussed the imminent arrival of GPT-5.6, a new update to the AI model behind ChatGPT. The startup also revealed that it would be the Trump administration that would approve use of the model on a case-by-case basis, on behalf of national security. This is why OpenAI was forced to agree to a phased rollout, contingent on Washington’s decisions. A

After the U.S. government suspended Anthropic’s latest Claude models for several weeks, it decided to tighten the screws on OpenAI. At the same time, the new GPT models were therefore reserved for a select group of handpicked insiders. The company had then described this restriction as a “short-term measure”, and said it did not want this kind of government control to become the norm.

GPT-5.6 Sol, along with Terra and Luna, will launch publicly this Thursday.

We’re expanding preview access globally now. pic.twitter.com/Uk5HcfSc2e

— OpenAI (@OpenAI) July 8, 2026

In a post published on X, OpenAI finally announced that the GPT-5.6 lineup would be rolled out worldwide starting Thursday, July 9. The startup apparently managed to meet Washington’s requirements. The models are no longer considered a threat to national security.

Also read: Why the big comeback of Claude Fable 5 disappoints everyone

OpenAI’s three new AIs

The lineup consists of three versions with distinct uses. The Sol model is the new flagship product from Sam Altman’s company. With enhanced capabilities in programming, science, and cybersecurity, it is presented as the most powerful GPT model to date. Meanwhile, Terra stands out for an excellent performance-to-cost ratio.AI delivers results close to GPT-5.5 at half the price. Finally, Luna is designed for tasks requiring efficiency and low latency. It is the fastest and most affordable option in the new family of AI developed by OpenAI.

According to the American media outlet Axios, the Trump administration ultimately gave the green light for large-scale deployment after a series of tests and meetings between OpenAI and government officials. Officials were concerned that the three AI models could be used to identify software vulnerabilities. OpenAI also classifies Sol, Terra and Luna in the “high” risk category for cybersecurity and biological and chemical risks. To assuage the authorities’ concerns and prevent abuse, the start-up has implemented a series of specific safeguards tailored to the profile of each model.

👉🏻 Follow tech news in real time: add 01net to your sources on Google, and subscribe to our WhatsApp channel.