CodeRabbit uses GPT-6.1 Sol to track down security risks in complex code
OpenAI has published that CodeRabbit uses its GPT-6.1 Sol model in a security agent that investigates vulnerabilities and proposes fixes to developers.

Finding a security flaw in a large software project is like looking for a leak in a hundred-story building: the problem is rarely where it shows. According to OpenAI, that is the territory CodeRabbit now works in, with help from GPT-6.1 Sol.
On October 26, 2026, OpenAI published a post on its website titled "GPT-6.1 Sol helps CodeRabbit strengthen security across complex codebases". Its summary states that CodeRabbit uses the model to investigate security risks in complex codebases, meaning the large, interconnected sets of files and dependencies that make up a software product.
At the center of it is a CodeRabbit security agent. An agent, in this context, is an AI system that doesn't just answer a question but carries out a task step by step, in this case investigating potential risks. According to the summary, the agent helps developers review the findings and the proposed fixes. That suggests the final say still rests with people: the tool flags issues and proposes fixes, and the team evaluates them.
For now, the picture is incomplete. It was not possible to read the full page, so this article relies only on the title and summary. There is no verified information on how the agent works, what types of vulnerabilities it detects, how accurately, or with what measurable results. It is also unknown whether GPT-6.1 Sol is generally available to developers, at what price, or whether CodeRabbit offers this feature to all of its customers or only some of them.
If those answers arrive, the case could serve as a reference point for development teams in Latin America that already use AI to review code and are wondering whether it can also take on part of the security work. Before drawing conclusions, it's worth reading the full announcement and CodeRabbit's documentation, and separating what OpenAI describes from what any team can actually use today.
Source: OpenAI