GPT-5.6 Debuts in US-Only Rollout as OpenAI Complies With Government Request
OpenAI has launched its GPT-5.6 series in a limited US-only preview after a request from the US government to adopt a phased rollout for advanced AI models. The new lineup includes Sol, Terra, and Luna, designed for high-performance, everyday, and cost-efficient AI tasks.
At the request of the US government, OpenAI began previewing its most recent powerful AI model series on June 26th with a select number of partners in the US. This news arrives two weeks after the US government caught Silicon Valley off guard.
On the grounds of national security, the Trump administration directed OpenAI competitor Anthropic to block access to its Fable 5 and Mythos 5 models for all foreign nationals. The GPT-5.6 series from OpenAI is making its debut with this version. This new series includes three models: Sol, the flagship of OpenAI; Terra, a mid-range model suitable for everyday tasks; and Luna, a quick and affordable alternative.
Why Trump Administration Banned the Launch?
The Mythos models by Anthropic and GPT-5.6 by OpenAI have both generated a lot of buzz because of their purportedly unparalleled capacity. Both AI versions possess the ability to detect software vulnerabilities, or openings in code that malicious actors can exploit. The White House's intervention against Anthropic stands in stark contrast to its previous efforts to reduce regulation of artificial intelligence (including attempts to prevent individual states from drafting their own regulations).
An executive order was passed earlier this month by Trump, establishing an optional federal examination of national security threats in sophisticated AI models before they are released. This move comes after pressure regarding Mythos' unique powers. In an effort to attract and retain clients in the face of intense competition from Anthropic and Google, Open AI announced that Terra would be priced half as much as its predecessor, GPT-5.5.
All the businesses receiving access to GPT 5.6 will be US-based, but OpenAI aims to include overseas partners next week. This is similar to Anthropic's US government fight over Mythos. However, the model will be accessible to employees of those organisations based abroad in "supported countries" (e.g., the UK and Australia). In order to fuel OpenAI's ChatGPT chatbot, the company had been collaborating with the US government on a model preview. Discussions with the National Cyber Director's Office and the Office of Science and Technology Policy led to the request for the staggered distribution.
Trump Administration Carving Framework for AI Models
A new voluntary framework to evaluate very advanced artificial intelligence models prior to their distribution was signed into law by the President of the United States this month. The President's previous deregulatory attitude on artificial intelligence has been reversed by this decree.
"Excessive regulation of the AI sector could kill a transformative industry," observed JD Vance, the Vice President, last year. The White House's position has shifted, though, in response to rapidly developing model capabilities, such as Mythos, which the UK's AI security body has praised as a "step up" from earlier state-of-the-art models.
