OpenAI has unveiled GPT-5-Codex, a new version of its AI coding agent designed to support software engineers with code reviews, flaw detection and multi-platform integration.
The tool follows earlier releases of Codex CLI and cloud-based Codex, expanding its functionality into terminal environments, IDEs such as Visual Studio Code, GitHub, cloud services and the ChatGPT mobile app. Subscribers to ChatGPT Plus, Pro, Business, Edu and Enterprise will have access to the new model.
According to OpenAI, GPT-5-Codex has been optimized to improve coding efficiency. The system can generate websites, provide detailed code comments and detect flaws, with the company claiming up to a 90% reduction in task completion times for certain cloud-based operations. “GPT-5-Codex can now run in the terminal, IDE software such as VS Code, in GitHub, on the cloud, and in the ChatGPT mobile app,” OpenAI said in its announcement.
The model operates in a sandboxed environment where internet and tool access are disabled by default. Developers can enable additional functions, including web access, allowing the AI to iterate projects and conduct code reviews independently.
Given its advanced capabilities, OpenAI has rated GPT-5-Codex as “high capability” in biological and chemical domains, activating additional safeguards to prevent misuse. The company said the classification reflects potential vulnerabilities that could allow the system to be manipulated for harmful purposes.
Source: Open AI