OpenAI Launches New macOS App for Agentic Coding
OpenAI has launched a new macOS app for Codex, introducing agentic workflows, multi-agent support and enhanced automation for AI-driven software development.
Topics
OpenAI has introduced a new macOS application for Codex, marking a significant step in its push into agentic coding workflows. The app integrates parallel multi-agent capabilities and advanced automation features designed to streamline AI-assisted software development.
The release builds on Codex’s earlier launches as a command-line tool and later a web interface. The macOS version brings together agent skills and state-of-the-art workflows in a more flexible desktop environment aimed at developers working on complex projects.
The launch follows the recent debut of GPT-5.2-Codex, OpenAI’s most advanced coding model to date. Chief Executive Officer Sam Altman said the new interface is intended to make the model’s advanced capabilities easier to use in real-world development scenarios.
The macOS app supports multiple AI agents operating in parallel and introduces background automations that can run on a scheduled basis. Tasks completed by agents are queued for review, enabling asynchronous workflows. Users can also customise agent behaviour by selecting personality styles ranging from pragmatic to empathetic.
Read More: IBM, Microsoft To Help Companies Accelerate Adoption of Gen AI
Benchmark comparisons show GPT-5.2-Codex performs competitively against rival systems, including models from Anthropic and Google. While results vary depending on the benchmark and use case, OpenAI is positioning Codex as a high-performance option for agentic development environments.
OpenAI said the broader goal is to accelerate software creation by reducing friction between idea generation and execution. The company believes that AI-assisted development tools can significantly shorten build cycles, allowing developers to create sophisticated applications in a fraction of the time required by traditional approaches.




































































































