OpenAI Releases GPT-5.2-Codex with Enhanced Cybersecurity Capabilities

On December 18, 2025, OpenAI introduced GPT-5.2-Codex, a sophisticated coding model tailored for software engineering and cybersecurity. This model enhances capabilities in long-horizon tasks, extensive code modifications, Windows features, and cybersecurity, although it does not achieve a 'High' rating according to OpenAI's Preparedness Framework. Compared to its predecessors, GPT-5-Codex and GPT-5.1-Codex-Max, it demonstrates notable improvements in cybersecurity assessments. It excels in performance on SWE-Bench Pro and Terminal-Bench 2.0, with improved abilities to analyze screenshots and diagrams. Currently accessible to paid ChatGPT users, API access is forthcoming. OpenAI is also testing a trusted access initiative for qualified professionals. Notably, on December 11, 2025, Andrew MacPherson identified vulnerabilities using GPT-5.1-Codex-Max.