OpenAI's Codex Hits 1 Million Active Users
OpenAI's code-generating AI, Codex, has reached one million active users, a major adoption milestone. The company now plans to make Codex the foundation for a wider push into business-facing AI agents that automate internal workflows. The surge suggests that AI-assisted development and prototyping are becoming core competencies for product managers, not just engineers.
OpenAI's Codex was first introduced in August 2021 as an AI model that could translate natural language into code. It was a fine-tuned version of GPT-3 and became the original engine that powered GitHub Copilot's popular code completion features in editors like VS Code. The initial Codex API was deprecated by OpenAI in March 2023, with GitHub Copilot transitioning to more advanced GPT-4 models. In May 2025, OpenAI relaunched the "Codex" brand, not as a simple code-completion model, but as a more autonomous software engineering agent built on its o3 reasoning architecture. This new version of Codex operates as a virtual software engineer within a secure, sandboxed cloud environment. It is designed to handle multi-step tasks such as writing entire features, running tests, fixing bugs, and proposing pull requests for review, moving far beyond its original function of just completing lines of code. The push into AI agents extends beyond coding and is central to OpenAI's enterprise strategy. The company is leveraging a platform called "Frontier" and forming alliances with major consulting firms like McKinsey, BCG, and Accenture to help businesses build and deploy AI agents that automate complex workflows across CRM, HR, and other internal systems. This positions OpenAI in a competitive landscape with other AI-assisted development tools. Key rivals include Amazon's CodeWhisperer, which is tuned for AWS services, and Google's Jules, an asynchronous coding agent powered by Gemini models. Access to the new Codex agent is available to users on ChatGPT's Pro, Team, and Enterprise plans. The push towards these more capable, autonomous agents reflects a broader industry shift from simple AI assistants to systems that can independently manage and execute complex professional tasks.
