OpenAI Launches ‘Codex’ – A Cloud-based Software Engineering Agent
Codex can perform tasks for you such as writing features, answering questions about your codebase, fixing bugs, and proposing pull requests for review; each task runs in its own cloud sandbox environment, preloaded with your repository.
OpenAI has launched Codex, a cloud-based AI software engineering agent integrated into ChatGPT.
Codex can perform tasks for you such as writing features, answering questions about your codebase, fixing bugs, and proposing pull requests for review; each task runs in its own cloud sandbox environment, preloaded with your repository.
Powered by the codex-1 model, a fine-tuned version of the o3 reasoning model, Codex excels in writing code, fixing bugs, running tests, and answering codebase queries in secure, sandboxed environments.
It achieved 75% accuracy in coding tests, surpassing the o3 model by 5%. Initially available to ChatGPT Pro ($200/month), Enterprise, and Team subscribers, access will expand to Plus and Edu users later.
Codex supports tasks taking 1–30 minutes, with customizable environments via AGENTS.md files. OpenAI emphasizes safety, training Codex to reject malicious requests, though manual code review is advised. The launch aims to boost developer productivity, competing with tools like Google’s Gemini Code Assist and Anthropic’s Claude, while addressing market demands amid potential acquisitions like Windsurf for $3 billion.
