OpenAI's new AI model is here, but not everyone can use it yet. The company has launched its GPT-5.6 series, which includes three models named Sol, Terra, and Luna, but only a small group of trusted partners in the US can access them for now.
This happened after the US government asked OpenAI to follow a safety check, much like the one Anthropic faced earlier with its Mythos and Fable models. OpenAI says this is just a short-term step and not something it wants to continue for long. To know more about OpenAI, GPT-5.6, US user access, and the new AI model rules, check out the full details below.
Why Did OpenAI Limit Access To Its New AI Model?
OpenAI said its new GPT-5.6 series went live on Friday, but only as a preview for a few partners in the US. This was not entirely OpenAI's decision and the company said it acted after the US government asked it to share details about what the new models can do. Even employees of these US partners working outside the country will get access, so the limit is more about which companies are involved than just location.
The Government Order Behind This Decision
Earlier this month, an order was signed asking AI companies to allow a voluntary safety check before launching strong new models. This came just two weeks after Anthropic was told to block all foreign users from its Mythos 5 and Fable 5 models. Anthropic shut down access completely, saying it could not follow that rule properly. OpenAI is facing a similar situation now, but it has been given a smaller, limited launch instead of a full block.
What Is Different About GPT-5.6
The GPT-5.6 series has three models made for different uses. Sol is the strongest and main model, then Terra is for everyday work and is expected to cost half of what its older version, GPT-5.5 charged once it is open to everyone. Luna is the fastest and cheapest option, and reports say these new models, like Anthropic's latest ones, are very good at finding problems in software code, which is one reason regulators are watching closely.