OpenAI has launched a new MacOS app for its Codex coding tool, marking a significant step in its effort to compete in the rapidly evolving space of agentic software development. As AI continues to reshape how code is written, with autonomous agents handling complex programming tasks, OpenAI is now bringing its most advanced coding capabilities into a dedicated desktop application tailored for developers. The new app builds on the evolution of Codex, which initially launched as a command-line tool in April and later expanded to a web interface. The MacOS version integrates multiple AI agents that can work in parallel, enabling sophisticated workflows where agents collaborate to plan, write, test, and debug code. This reflects a broader industry shift toward agentic development, a model already popularized by tools like Claude Code and Cowork. The launch arrives just weeks after OpenAI introduced GPT-5.2-Codex, its most powerful coding model to date. CEO Sam Altman emphasized the model’s capabilities during a press call, stating, “If you really want to do sophisticated work on something complex, 5.2 is the strongest model by far.” However, he acknowledged that usability had been a challenge. “Taking that level of model capability and putting it in a more flexible interface, we think is going to matter quite a bit.” While GPT-5.2 leads on TerminalBench—a benchmark measuring command-line coding performance—results from other tests tell a more nuanced story. Gemini 3 and Claude Opus have posted scores within the margin of error, suggesting strong competition. Similarly, on SWE-bench, which evaluates AI’s ability to fix real-world software bugs, no single model has established a decisive lead. Despite these mixed benchmark results, OpenAI is betting on user experience and workflow innovation. The new Codex app introduces several features designed to enhance productivity. Users can now set up automated workflows that run in the background on a schedule, with results queued for review later. The app also lets developers choose different agent personalities—ranging from pragmatic to empathetic—tailoring the AI’s tone and approach to match individual working styles. For OpenAI, the ultimate promise lies in speed. “You can use this from a clean sheet of paper, brand new, to make a really quite sophisticated piece of software in a few hours,” Altman said. “As fast as I can type in new ideas, that is the limit of what can get built.” The company sees the app not just as a coding assistant, but as a full development partner that accelerates innovation from concept to deployment.