OpenAI launches new macOS app for proxy encryption

AI has already had a seismic impact on how software is written, with much of the arduous programming work now being done by swarms of agents and sub-agents. But as developers experiment with new interfaces and form factors for human-AI collaboration, it’s becoming harder for even the most advanced AI labs to keep up.

The current trend is the development of agentic software – systems in which AI agents can work autonomously on programming tasks – exemplified by the Claude Code and Cowork applications. Meanwhile, OpenAI is gradually building out its own Codex tool, which it launched as a command-line tool last April and expanded to a web interface a month later.

Now OpenAI is taking a big step towards catching up. The company on Monday launched a new macOS app for Codex, integrating many of the proxy practices that became popular in the past year. The new app is designed to work with multiple agents in parallel, integrating agent skills and other cutting-edge workflows. The launch also comes less than two months after the launch of GPT-5.2-Codex, OpenAI’s strongest encryption model, which the company hopes will be enough to entice Claude Code users.

“If you really want to do cutting-edge work on something complex, 5.2 is the most powerful model ever,” Sam Altman, the company’s CEO, told reporters on a press call. “However, it was more difficult to use, so, by taking that level of model capability and putting it into a more flexible interface, we think it will be quite important.”

While Altman’s confidence in GPT-5.2 is understandable, the coding standards tell a more complex story. GPT-5.2 holds the top spot in TerminalBench (a test that measures how well an AI handles command-line programming tasks), at least as of press time. But the clients from Gemini 3 and Cloud Opus scored almost equally — lower, but within the benchmark’s margin of error. Results from SWE-bench, another coding benchmark that tests AI’s ability to fix software errors in the real world, are similar, and show no clear advantage for GPT-5.2. However, agent use cases have been difficult to measure effectively, and modern models can vary greatly in user experience.

The Codex app also comes with a host of new features that OpenAI says will help it achieve parity, or in some cases, outperform Claude’s various apps. Codex will allow automations that can be set to run in the background on an automated schedule, with the results queued for review when the user returns. Users can also choose different agent personalities – from realistic to sympathetic – depending on their work style.

But for the company, the biggest selling point is the sheer speed of development made possible by AI. “You can use this from a clean sheet of paper, brand new, to create a very sophisticated program in just a few hours,” Altman said. “As much as I can write new ideas, this is the limit of what can be built.”