OpenAI has launched a research preview of its latest AI tool for software development—Codex. It is a cloud-based engineering agent capable of independently performing multiple programming tasks. Codex is now available for ChatGPT Pro, Team, and Enterprise users, and is positioned as a significant step in OpenAI’s strategy to expand deeper into the fast-evolving AI-assisted coding landscape. It is built on codex-1, a variant of OpenAI’s o3 model customized for software engineering.
The new agent can write new code, identify and fix bugs, execute test cases, and answer queries about existing codebases. What sets Codex apart is its ability to function autonomously in cloud sandboxes while maintaining access to a developer’s repository, allowing for parallel task execution without direct human input during runtime.
OpenAI’s engineering lead, Srinivas Narayanan, described Codex as a “fundamentally new way of working,” emphasising its potential to shift how developers interact with code. Unlike current co-pilots that assist in real time, Codex operates independently and returns results in one to 30 minutes, depending on task complexity.
The release comes amid growing competition in AI coding tools, with major players like Microsoft, Google, Amazon, and startups like Anysphere, which developed Cursor, heavily investing in the space. OpenAI is also reportedly in talks to acquire Windsurf, an AI coding startup, for around $3 billion, though both companies have declined to comment on the speculation.
See What’s Next in Tech With the Fast Forward Newsletter
Tweets From @varindiamag
Nothing to see here - yet
When they Tweet, their Tweets will show up here.
