OpenAI on Friday rolled out a powerful new tool for software developers, as the company pushes further into automating coding tasks with AI.
The new product, called Codex, is an AI agent designed to help programmers write code, fix bugs, and run tests — often simultaneously.
“Technical teams at OpenAI have started using Codex as part of their daily toolkit,” OpenAI said in a blogpost. “It is most often used by OpenAI engineers to offload repetitive, well-scoped tasks, like refactoring, renaming, and writing tests, that would otherwise break focus.”
“When uncertain or faced with test failures, the Codex agent explicitly communicates these issues, enabling users to make informed decisions about how to proceed,” OpenAI added.
Unlike traditional chatbots that respond to prompts and generate responses in mostly words, AI agents like Codex can interact with other software and online services, such as helping you with a DoorDash order or booking a dinner reservation.
The Codex rollout came after OpenAI launched GPT-4.5 in February. A livestream demo highlighted its improved reasoning, intuition, and reduced hallucinations.
CEO Sam Altman described it as “the first model that feels like talking to a thoughtful person,” but also said that its intelligence and nuance comes at steep computational cost. Due to GPU shortages, GPT-4.5 was initially available only to $200-per-month ChatGPT Pro users.
Codex is now available to subscribers of OpenAI’s ChatGPT Pro plan, which costs $200 a month. OpenAI also said it will eventually bring Codex to its other premium offerings.
OpenAI did not immediately respond to a request for comments.
