OpenAI has spent two years telling people that ChatGPT can answer almost anything. With its latest release, it wants ChatGPT to finish the job instead. The company has introduced ChatGPT Work, an agent inside ChatGPT that it says can take action across a user’s apps and files, stay with a project for hours if needed, and turn a goal into completed work.
According to OpenAI, the agent can gather information across a person’s apps and workflows to produce finished materials such as sheets, slides, docs and web apps, breaking complex projects into smaller steps it completes on its own.
It is powered by GPT-5.6, the frontier model that launched the same day as part of a broad rollout that cleared US government testing, which OpenAI says makes ChatGPT state of the art at reasoning through multi-step tasks and at producing materials that follow a user’s own templates and reference files.
Codex is the engine underneath. With Codex technology built in, OpenAI says ChatGPT can now get real work done across web, mobile and desktop, and it notes that more than five million people use Codex every week, with over a million now using it for tasks outside software development. That lineage extends the company’s decision to merge ChatGPT and Codex under co-founder Greg Brockman.
