OpenAI's New Codex Update Stacks It Up For (Almost) Everything, As It Challenges Anthropic

OpenAI is stepping up its push into AI-powered coding with a sweeping overhaul of Codex, aiming to close the gap with rivals like Anthropic, whose Claude Code has recently gained strong traction among enterprise users.

The company introduced a major update to Codex, transforming it from a coding assistant into a broader, more autonomous development partner. The move comes amid intensifying competition with Anthropic, where Claude Code is increasingly becoming the preferred choice for businesses.
 

TL;DR

 

• OpenAI upgrades Codex with deeper automation and workflow integration
• Tool can now operate computers, use apps, and handle multi-step tasks
• Adds memory, scheduling, and proactive task suggestions
• Expands plugins and supports full development lifecycle workflows
• Aims to counter Anthropic’s growing lead in AI coding tools

 

Codex Expands Beyond Coding Into Full Workflow Automation

The latest iteration of Codex pushes it well beyond writing code. It can now interact directly with a user’s computer, navigating apps, clicking, typing, and executing tasks through its own interface.

This enables developers to automate workflows that previously required manual intervention, especially in environments where APIs are unavailable. Codex can also run multiple agents in parallel, allowing simultaneous task execution without interrupting the user’s ongoing work.

In addition, the tool now includes an in-app browser, letting developers guide tasks visually by commenting directly on web pages. This is particularly useful for frontend development and iterative design work.

 

New Features Bring End-To-End Software Development Support

OpenAI is positioning Codex as a unified workspace for the entire software development lifecycle. The update introduces features like pull request reviews, multi-terminal support, and remote development via SSH.

Developers can now manage files, preview documents, and track project progress through a centralized interface. A new summary pane provides visibility into agent actions, sources, and outputs, helping users stay aligned with ongoing tasks.

The addition of over 90 plugins further extends Codex’s reach. Integrations with tools like GitLab, Microsoft services, and project management platforms enable it to gather context and act across a wide range of developer environments.
 

AI Memory And Automation Push Codex Toward Persistent Intelligence

One of the most significant updates is Codex’s ability to retain memory and build context over time. It can now remember user preferences, learn from previous interactions, and apply that knowledge to future tasks.

The platform also introduces automation capabilities that allow it to schedule and resume tasks independently. This means Codex can continue working on long-term projects across days or even weeks without constant input.

It can even proactively suggest next steps, identifying pending tasks from connected tools like Slack, Gmail, or project documents, and presenting prioritized actions to users.
   

OpenAI Aims To Reclaim Ground In Intensifying AI Coding Race

This overhaul reflects OpenAI’s broader strategy to regain momentum in the AI coding space. With Anthropic’s Claude Code gaining enterprise traction, Codex’s transformation into a more autonomous, context-aware system signals a direct response.

By combining coding, automation, and intelligent task management into a single platform, OpenAI is attempting to redefine what AI-assisted development looks like.

The company’s long-term ambition remains clear, narrowing the gap between what developers envision and what they can build, while embedding AI deeper into everyday workflows