OpenAI GPT-5.2-Codex Supercharges Agentic Coding and Vulnerability Detection
OpenAI has unveiled GPT-5.2-Codex, a cutting-edge model optimized for agentic coding and enhanced cybersecurity tasks. The release highlights breakthroughs in handling complex software engineering and vulnerability detection. GPT-5.2-Codex tops SWE-Bench Pro with 56.4% accuracy, outperforming GPT-5.2 at 55.6% and GPT-5.1 at 50.8%. On Terminal-Bench 2.0, it scores 64.0%, surpassing prior versions like GPT-5.2’s 62.2%. These […]
The post OpenAI GPT-5.2-Codex Supercharges Agentic Coding and Vulnerability Detection appeared first on Cyber Security News.