Skip to content
agentusecases

Codex use cases

OpenAI · Coding agent · 1 use case

OpenAI's coding agent that works in the terminal, the editor and the cloud: it reads the repository, edits code and runs the tests.

Visit Codex

1 use case

  • Run thousands of Codex agents in parallel on 36- to 48-hour quant research sessions

    The startup Third Moment Research runs up to 10,000 agents at once, each using the default Codex harness in a Docker environment with 1 TB of market data and a research repository, to improve a baseline trading system. An outer harness examines tasks left unfinished and decides whether to retry, terminate or escalate to human review. By analyzing thousands of agent trajectories, Codex found a backtest bottleneck and shipped the fix to production. We wrote the prompt below, since the source doesn't share the one used.

    Codex

About Codex

Codex is OpenAI's coding agent. It runs in the terminal, the editor and the cloud, handling several tasks in parallel: it reads the repository, proposes changes, runs the tests and hands back a diff or pull request for review.

In this directory, Codex appears in 1 use case, mostly in Research & Analysis. The latest one was added on Oct 14, 2026.

Where it shows up well

  • Parallel coding tasks in the cloud
  • Fixing bugs from a description
  • Writing and running tests
  • Reviewing pull requests

Other agents