OpenAI Debuts GPT-5.2 Codex, Pushing Autonomous Coding and Cybersecurity Forward

Details

• OpenAI has introduced GPT-5.2-Codex, a code-focused version of its latest AI model, now available through the Codex API and Playground.
• This update highlights autonomous workflows, enabling the AI to independently plan, write, test, and refactor entire multi-file repositories without heavy user input.
• The new model offers enhanced reliability on complex coding tasks and sustains its performance when operating at enterprise scale, addressing earlier limitations.
• A newly integrated cybersecurity suite allows for automated vulnerability scanning, exploit reproduction, and patch creation in languages such as C, Rust, and Solidity.
• Current GPT-5.1-Codex users are automatically upgraded, and broader access for new users begins December 19, with pricing remaining unchanged.

Impact

This release heightens competition with solutions like Anthropic’s Claude-Code and Google’s Gemini-Code-Gem, both of which recently introduced their own autonomous features. By holding pricing steady, OpenAI could drive more businesses to adopt AI-powered, secure development practices, challenging traditional tools. The product’s cybersecurity emphasis is likely to appeal to enterprises navigating new regulations and seeking integrated AI dev platforms.