There is currently a low-grade war between OpenAI and Anthropic over who can release the most convenient and powerful AI-coding tools. While Anthropic’s Claude Code has been the tool of choice for many businesses, OpenAI isn’t giving up yet.

This week, OpenAI announced a revamp of Codex with new features that allow it to operate in the background on your computer, opening any app and carrying out operations that can be as simple as clicking and typing. This agentic update means users can still work on their own projects while Codex handles auxiliary tasks like iterating on frontend changes or testing apps.

Other updates include an in-app browser to issue commands, a new memory feature allowing Codex to recall previous work sessions, and image-generation abilities. Additionally, 111 plug-in integrations from popular tools such as CodeRabbit and GitLab Issues have been announced, giving Codex the ability to carry out minor clerical tasks.

The way OpenAI has framed it, these updates are aimed at making Codex a competitive coding assistant that can be integrated into various corporate workflows. However, some of the powers added seem similar to those previously released by Anthropic for Claude Code, raising questions about the nature of competition in this space.