OpenAI has introduced a broad set of updates to Codex, adding background agent capabilities, browser control, and enterprise integrations, as competition with Anthropic intensifies in the market for AI coding tools, where Claude Code has recently gained traction among business users.
Competition With Anthropic Shapes Product Direction
A growing rivalry between OpenAI and Anthropic centers on delivering more capable and convenient AI coding systems. Claude Code has emerged as a preferred option for some businesses, according to prior reporting, placing pressure on OpenAI to expand Codex’s functionality.
Recent updates indicate a shift toward broader enterprise use cases, as OpenAI positions Codex as both a coding assistant and a tool that can integrate into wider workflows.
Background Agent Enables Parallel Task Execution
The most notable addition allows Codex to operate in the background on a user’s computer. The system can open desktop applications and perform actions using a cursor that clicks and types.
This setup enables multiple agents to run in parallel on a Mac, according to OpenAI, without interrupting active user work. The agents handle auxiliary coding-related tasks while the user continues primary work in other applications. OpenAI lists use cases such as iterating on frontend changes, testing applications, and interacting with software that does not provide an API.
New Features Mirror Capabilities Seen In Claude Code
Some of the added capabilities resemble features previously introduced by Anthropic. The company had announced that its tools, including Claude and Cowork, could remotely control a user’s desktop to perform tasks without direct input.
The overlap highlights how both companies are developing agent-based systems that extend beyond code generation into direct system interaction.
Browser Integration Expands Web-Based Task Handling
Codex now includes an in-app browser that allows users to issue commands for execution within web applications. OpenAI states that this feature supports workflows such as frontend and game development.
The company indicates plans to extend this capability further, aiming for Codex to command browser functions beyond locally hosted web applications.
Memory And Image Generation Add Context And Output Capabilities
A preview feature called “memory” enables Codex to retain information from previous sessions and build context around user workflows. This allows the system to adapt responses based on prior interactions.
The tool also gains image-generation functionality. OpenAI says this can support tasks such as creating product concepts, slide visuals, mockups, and placeholder images.
Plugin Ecosystem Connects Codex To External Tools
OpenAI has introduced 111 plug-in integrations, including connections to CodeRabbit and GitLab Issues. These integrations allow Codex to perform actions within external platforms.
The company describes these capabilities as enabling clerical and organizational tasks. For example, Codex can review activity across communication tools and calendars to generate daily task lists.
Pricing Update Targets Enterprise Flexibility
A new pay-as-you-go pricing model for Codex has been introduced for ChatGPT enterprise and business users. The option is intended to provide more flexibility in how organizations adopt and scale usage of the tool.
Ongoing Strategic Shift And Industry Pressures
OpenAI has recently adjusted its focus toward enterprise offerings while reducing emphasis on certain consumer-facing projects, including its social video app Sora 2. At the same time, the company faces ongoing scrutiny, including lawsuits related to the alleged mental health impact of ChatGPT on some users.
Featured image credits: Wikimedia Commons
For more stories like it, click the +Follow button at the top of this page to follow us.