OpenAI recently launched its new AI coding agent Codex, which is now integrated into ChatGPT. It is a cloud-based software engineering tool that can handle multiple development tasks simultaneously. Codex is specifically designed to help with software development and is powered by Codex-1, a variant of OpenAI's high-tech O3 reasoning model. Let's know 5 special things about it...
1. What is Codex?
OpenAI announced its new AI coding agent Codex in a live stream on Friday. It is a coding assistant built into ChatGPT that can handle multiple tasks simultaneously. It can perform tasks such as feature implementation, bug fixing, and providing technical information related to the codebase. Each task runs in a secure, isolated sandbox environment, which is already connected to your project's repository.
2. What technology is it based on?
Codex is based on codex-1, an optimized version of OpenAI's o3 reasoning model. This model is trained on real-world software development scenarios so that it adopts a human-like coding style, follows instructions correctly, and keeps improving the code until the test passes.
3. For which users is it available?
Codex is currently being rolled out in a phased manner to users with ChatGPT Pro, Enterprise, and Team plans. It will soon be included in ChatGPT Plus and Edu plans as well, although no specific date has been given for this.
4. How to use?
Codex is available in the sidebar of ChatGPT. There, you can enter a prompt and select “Code” to assign a new programming task. If you have any questions about the codebase, you can ask with the “Ask” button. Each task is completed in a separate cloud workspace with your code files pre-loaded.
5. What happens after the task is completed?
When Codex completes a task, it saves the changes to its sandbox. It also documents the entire process via terminal logs and test results so users understand what was done. You can then review the changes, create a Pull Request on GitHub, or apply the updates directly to your development files.
Disclaimer: This content has been sourced and edited from Amar Ujala. While we have made modifications for clarity and presentation, the original content belongs to its respective authors and website. We do not claim ownership of the content.