OpenAI Takes Codex Coding Agent to General Availability

OpenAI has moved Codex, its AI coding agent, from research preview to general availability, and shipped three new features alongside the release: a Slack integration, a software development kit for embedding the Codex agent into custom workflows, and administrative tools for managing Codex at scale across ChatGPT workspaces.

The announcement lands amid rapid adoption. According to OpenAI, daily usage of Codex has grown by more than 10x since early August, and GPT-5-Codex, the model purpose-built for the agent, has served over 40 trillion tokens in the three weeks since its launch. OpenAI calls it "one of our fastest growing models ever."

The stakes are significant. AI coding assistants have become the fastest-moving category in enterprise AI adoption, with GitHub Copilot, Anthropic's Claude Code, and Google's Gemini Code Assist all competing for developer workflows. Codex general availability signals OpenAI's push to move the product from experiment to default infrastructure for engineering teams — a shift visible inside the company itself.

"Nearly all engineers use Codex today, up from just over half in July," OpenAI states. Those engineers merge 70% more pull requests each week, and Codex automatically reviews almost every PR to catch critical issues before they reach production.

Customers span from startups to large enterprises. OpenAI names Duolingo and Vanta on the startup side, and Cisco and Rakuten among large enterprises using Codex.

Codex in Slack

The new Slack integration lets developers delegate tasks to Codex without leaving a conversation. Tag @Codex in a channel or thread, and it will automatically gather the context it needs from the conversation, choose the right environment, and answer with a link to the completed task in Codex cloud. From there, users can merge Codex's changes, keep iterating, or pull the task to their computer to continue working locally.

The framing is deliberately human: OpenAI describes it as delegating tasks or asking questions "just like you would a coworker." Setup instructions are available in OpenAI's developer docs.

Codex SDK

The Codex SDK is the more strategically significant release. It lets engineering teams embed the same agent that powers the Codex CLI into their own workflows, tools, and apps — with what OpenAI claims is "state-of-the-art performance on GPT-5-Codex without extra tuning."

GPT-5-Codex was trained specifically for the open-source agent implementation behind the Codex CLI, and OpenAI says it tuned the agent implementation so that its prompt, tool definitions, and agent loop deliver faster and more accurate results with models like GPT-5-Codex.

The SDK ships with structured outputs for parsing agent responses and built-in context management to resume sessions. It is available for TypeScript today, with more languages coming soon. OpenAI is also releasing a new GitHub Action for integrating Codex into CI/CD pipelines, and developers running workflows in shell environments can install and run the Codex CLI directly with codex exec.

Admin features

For enterprises, the general availability release adds controls aimed at IT and security teams. ChatGPT workspace admins can now edit or delete Codex cloud environments — for example, to remove sensitive information or clean up unused environments. They can enforce safer defaults for local usage with the Codex CLI and IDE extension, define overrides through managed configuration, and monitor actions taken by Codex.

New analytics dashboards let admins track usage across the CLI, IDE, and web, along with the quality of code reviews Codex provides. The dashboard includes charts for daily code review issues by priority and sentiment of code review feedback over time.

Codex at work

OpenAI shared two customer examples that illustrate where the product is landing.

At Cisco, engineers use Codex to speed up review of complex pull requests, reducing review times by up to 50%. Codex reviews both human-written code and its own output, which Cisco says helps engineers maintain high standards while enabling faster product releases.

At Instacart, the Codex SDK is integrated with Olive, the company's background coding agent platform. Engineers spin up a remote development environment and complete end-to-end tasks with a single click, using Codex to edit and test changes. Codex automatically cleans up tech debt such as dead code and expired experiments, improving code quality and reducing latency across codebases. It also handles repetitive, well-understood changes, reducing backlog and accelerating engineering velocity.

Availability and pricing

The Slack integration and Codex SDK are available starting today to developers on ChatGPT Plus, Pro, Business, Edu, and Enterprise plans. The new admin features are limited to Business, Edu, and Enterprise customers.

One pricing change is coming: starting October 20, Codex cloud tasks will begin counting toward Codex usage. OpenAI has published a pricing breakdown for each plan in its help center.

The combination — general availability, an SDK, Slack access, and enterprise-grade admin controls — positions Codex to compete not just as a coding assistant but as an embeddable agent platform, and the October 20 usage-accounting change will be the first real test of how much developers value it when the meter starts running.