OpenAI released GPT-5.6 after government approval and introduced ChatGPT Work
OpenAI received approval from the Trump administration and opened public access to GPT-5.6, which had previously been tested only by organizations authorized by the authorities. At the same time, the company introduced ChatGPT Work, an agent for non-technical tasks that combines ChatGPT and Codex. The new product runs on the Sol, Terra, and Luna model families.
AI-processed from The Verge; edited by Hamidun News
OpenAI received approval from the Trump administration on February 10, 2026 to publicly launch the GPT-5.6 model, which had been tested for about two weeks prior by only government-authorized organizations. On the same day, the company unveiled ChatGPT Work — an AI agent for everyday non-technical tasks based on Codex and the GPT-5.6 family.
Why the launch was delayed
The public release of GPT-5.6 turned out to be tied to regulatory constraints. During the limited preview phase, the model was provided only to organizations that received permission from the U.S. government. This testing format lasted approximately two weeks until OpenAI received approval for a broader rollout.
The Trump administration's decision removed a key obstacle to the public distribution of the model. The original announcement does not specify what requirements were imposed on GPT-5.6 or which organizations participated in the limited testing.
- The public launch received approval from the Trump administration
- The limited preview lasted approximately two weeks
- Before approval, the model was available only to government-authorized organizations
- OpenAI CEO Sam Altman called GPT-5.6 the company's best model
"This is the best model we have ever created," said
Sam Altman.
What ChatGPT Work can do
ChatGPT Work combines the capabilities of ChatGPT and Codex to make code-working tools useful for people without technical training. The agent is not designed only for programming: OpenAI positions it as a tool for performing non-technical tasks.
The presented description does not provide a complete list of ChatGPT Work's scenarios and features. It is known only that the agent leverages Codex capabilities beyond direct software development and operates on the GPT-5.6 model family.
The new system includes three models — Sol, Terra, and Luna. The original material does not specify how they differ from each other, what tasks each model solves, or whether they are available to users individually.
What it means
OpenAI is simultaneously expanding access to its new flagship model and transforming Codex from a tool primarily for developers into the foundation of a more universal work agent. Regulatory approval also demonstrates that for public launches of advanced models, the company increasingly needs to account for the position of government agencies.
For users, the key takeaway is that GPT-5.6 is coming out of limited preview, and ChatGPT Work offers the ability to use Codex capabilities in tasks that do not require code writing. At the same time, OpenAI has disclosed few technical details about the model, the agent, and the distribution of roles among Sol, Terra, and Luna.
Need AI working inside your business — not just in your newsfeed?
I build production AI for companies — custom CRM, internal tools, autonomous agents, workflow automation. Owned by you, shaped to your process, no per-seat tax. Built by Zhemal Khamidun, CPO of AlpinaGPT (AI platform, 6,000+ users).
The AI world, distilled — once a week
Seven stories that actually mattered, hand-picked. No noise, no reposts, no press releases.
Done! Check your inbox for a confirmation.