OpenAI launches GPT-5.6, the new engine behind ChatGPT and Copilot

What happened?

OpenAI released the GPT-5.6 family of models: Sol, the new flagship, alongside Terra for everyday work and Luna, the budget tier. The rollout started July 9 and reaches ChatGPT users over the following days, with free users getting Terra. OpenAI also announced that GPT-5.6 is now the preferred model for Microsoft 365 Copilot, the AI inside Word, Excel, and PowerPoint. The reported gains are largest in agentic browsing and long professional tasks, which is the machinery ChatGPT uses to search the web and assemble answers.

Why does it matter?

When the engine under ChatGPT changes, the answers can change with it. Which pages get found, how they are summarized, and who gets named as a source can all shift, even though OpenAI has not described specific changes to citation behavior. Through Copilot, the same engine now sits inside the office software many organizations use every day. Any visibility baseline you measured last month was measured against a model that is no longer the one answering.

Should I do anything?

Nothing urgent, and no content changes are needed on the strength of a model release alone. But this is the right week to re-run your visibility spot check: ask ChatGPT the questions your most important pages should answer and compare against your last results. If you keep a simple log of these checks, with the date, the question, and whether you were cited, a model release is exactly when it pays off. If you were cited before and are not now, start with our audit guide below.

Learn more

• OpenAI: GPT-5.6 announcement
• OpenAI: GPT-5.6 in Microsoft 365 Copilot
• TechCrunch on the Microsoft angle
• Our guide: conduct an AI visibility audit