Photo of DeepakNess DeepakNess

OpenAI launches Codex

Unproofread notes

OpenAI launches Codex, a cloud-based agent that writes code and works on multiple tasks at once. It's just launched, and can be accessed from inside ChatGPT at chatgpt.com/codex but visiting this URL just redirected me back to ChatGPT as it's only for ChatGPT Pro users, and not Plus users.

Currently, it's in a research preview but it's said to have features like:

The implementation is very interesting as it runs in its own cloud sandbox environment, and can be directly connected to your GitHub repo. It performs better than o1-high, o4-mini-high, and o3-high.

The cool thing is, it can also be guided by an AGENTS.md file placed within the repository. Very cool.

Today, we’re also releasing a smaller version of codex-1, a version of o4-mini designed specifically for use in Codex CLI.

Yes, they're also releasing something for Codex CLI as well. And about the pricing and availability:

Starting today, we’re rolling out Codex to ChatGPT Pro, Enterprise, and Team users globally, with support for Plus and Edu coming soon. [...] We plan to expand access to Plus and Edu users soon.

For developers building with codex-mini-latest, the model is available on the Responses API and priced at $1.50 per 1M input tokens and $6 per 1M output tokens, with a 75% prompt caching discount.

I am excited to see how this compares to Claude 3.7 Sonnet and Gemini 2.5 Pro in terms of coding, fixing bugs, designing UI, etc. I also uploaded a quick video about the same that you can watch on YouTube.

Comment via email