GPT-5-Codex at a glance
GPT-5-Codex adjusts its thinking time to the task, worked independently for more than 7 hours in OpenAI's tests, and was trained for code review across the Codex CLI, IDE extension, cloud and GitHub.
What changed with GPT-5-Codex
On September 15, 2025, OpenAI released GPT-5-Codex, a version of GPT-5 further optimized for agentic software engineering in its Codex product. It was trained on real engineering work such as building projects from scratch, adding features and tests, debugging and large refactors.
OpenAI said the model spends less time on small requests and more on complex ones, and that in testing it worked on its own for over 7 hours on large tasks, fixing test failures until it succeeded. It was also trained to review code by running the code and tests to validate changes. GPT-5-Codex powers Codex in the CLI, the IDE extension, the cloud and GitHub, included with ChatGPT Plus, Pro, Business, Edu and Enterprise; an update on September 23 added API-key access at GPT-5 pricing.
Where to compare GPT-5-Codex
Independent leaderboards and official resources for comparing this model yourself. Rankings change often; each site shows its own current data.
General leaderboards
Why GPT-5-Codex matters for businesses
For businesses with software teams, a coding agent that can take a ticket, work for hours and review pull requests changes how small teams staff projects, but its output still needs human review, tests and security checks before deployment.
Sources for this story
- Introducing upgrades to Codex (opens in a new tab)Official
OpenAI · openai.com
- GPT-5 Codex (high) model page (opens in a new tab)Benchmark
Artificial Analysis · artificialanalysis.ai
Last verified: . IA Compa summarizes public announcements for business readers. We are not affiliated with the companies mentioned. Always confirm details, pricing, and availability with the official source.
- #llm
- #openai
- #gpt-5
- #coding
- #agents
