Over a month ago, OpenAI introduced a series of GPT-5.2 models, including GPT-5.2-Codex, an agent programming model for complex software engineering in real-world scenarios. The model was made available via the OpenAI API a few weeks ago, and now it’s also available on GitHub Copilot.
Over a month ago, OpenAI introduced a series of GPT-5.2 models, including GPT-5.2-Codex, an agent programming model for complex software engineering in real-world scenarios. The model was made available via the OpenAI API a few weeks ago, and now it’s also available on GitHub Copilot.
The GPT-5.2-Codex model is now available to users with Copilot Enterprise, Copilot Business, Copilot Pro, and Copilot Pro+ subscriptions, Neowin writes .
This means that you can access the model through GitHub Copilot, integrated into various IDEs, including:
Visual Studio versions 17.14.19 and later, as well as 18.0.0 and later, in all modes: ask, agent;
GitHub Copilot for JetBrains version 1.5.61 and later, in all modes: ask, edit, agent;
GitHub Copilot for Xcode version 0.45.0 and later, in all modes: ask, agent;
GitHub Copilot for Eclipse version 0.13.0 and later, in all modes: ask, agent.
In addition to the above-mentioned IDEs, the model will be accessible via GitHub Copilot Chat on the GitHub website and in the GitHub Mobile app.
However, system administrators with access to Enterprise and Business subscriptions must first enable the GPT-5.2-Codex policy and provide consent in Copilot settings. Once this is done, users in the organization will be able to select the new model in the model selection interface in their respective IDEs.
GPT‑5.2-Codex is a version of GPT‑5.2 further optimized for agent development in Codex, including improvements for long-term performance through context compression, better performance during large code changes such as refactorings and migrations, improved performance in Windows environments, and significantly stronger cybersecurity capabilities.
