OpenAI Codex
OpenAI's software-engineering agent: delegate whole coding tasks to cloud environments that implement, test, and iterate — in parallel while you do other work.
OpenAI's software-engineering agent: delegate whole coding tasks to cloud environments that implement, test, and iterate — in parallel while you do other work.
Codex is built for delegation. Instead of pairing line-by-line, you hand it a task — a feature, a bug fix, a migration — and it works in its own cloud environment: writing code, running tests, and iterating.
Because tasks run in the cloud, you can spin up several in parallel. It's a different rhythm from terminal agents: less pairing, more assigning and reviewing.
Run multiple coding tasks simultaneously in isolated environments.
Hand off features or fixes and review the finished diff.
Runs test suites and revises until things pass.
Drive it from the terminal as well as the web interface.
Developers and teams who want to delegate well-defined tasks — bug fixes, migrations, test coverage — and review finished work in batches.
Pricing and features change fast — check the official site for the latest.
Visit official site →Last updated: October 2026 · Compare all agents