OpenAI Group PBC has launched GPT-5.2-Codex, an advanced artificial intelligence coding model that automates complex software engineering tasks. Building on the capabilities of GPT-5.2, this version enhances context compaction, large code refactoring, performance in Windows environments, and cybersecurity features. Achieving a remarkable 56.4% accuracy on the SWE-Bench Pro benchmark and 64% on Terminal-Bench 2.0, GPT-5.2-Codex sets a new standard among coding models. Its strengthened vision capabilities enable interpretation of screenshots and technical diagrams, facilitating the transformation of software design mockups into functional prototypes.
Designed to improve software quality and maintainability, GPT-5.2-Codex excels in refactoring, helping reduce memory usage and increase response times. This latest release enhances security, critical for AI-driven software engineering, enabling teams to identify and mitigate vulnerabilities effectively. Available now for paid ChatGPT users, its capabilities promise increased efficiency in coding and a competitive edge in software development.
Source link