OpenAI's AI agents reach 10 million users after ChatGPT Work debut

OpenAI has announced that 10 million people are now using its two agent products -- Codex for coding and ChatGPT Work for broader multi-step tasks -- nearly doubling usage from earlier in July 2026. The milestone came less than two weeks after the launch of ChatGPT Work, OpenAI's push to turn ChatGPT from a chatbot into a tool that finishes jobs rather than just describing them.

What happened

OpenAI will announce that its agentic products have reached 10 million weekly active users combined, according to a Bloomberg report. Codex passed 5 million weekly users at the end of May, reached 6 million on July 12, and roughly 8 million by mid-July. More than a million of those users now apply Codex to work outside software development. Usage of the agentic tools as a whole rose about 2.5 times in the week after OpenAI shipped GPT-5.6, the model that powers ChatGPT Work.

ChatGPT Work launched on July 9 for Pro, Enterprise, and Edu subscribers, then expanded to Plus and Business plans within days. Rather than answering a prompt and stopping, it splits a brief into smaller steps and works through them, generating spreadsheets, slide decks, documents, and small web apps while staying on a project for hours at a time. The agent connects to services such as Slack and Google Drive, can schedule recurring tasks, and uses OpenAI's Computer Use feature to operate desktop software on a user's behalf.

Why AI builders should care

For teams building AI products and workflows, the rapid adoption of ChatGPT Work signals that enterprise buyers are ready to trust agents with real work -- not just drafting text but executing multi-step processes across apps and files. The usage-based pricing model, already applied to Codex, gives administrators on Enterprise and Edu plans the ability to cap spending, restrict which tools the agent can touch, and review activity through a compliance interface. This makes it easier for IT leaders to approve agent deployments without losing control.

The growth also puts pressure on competitors. Google recently launched Gemini Spark, an always-on assistant that takes instructions through Gmail, while Anthropic has pushed its own coding and workplace agents. All are chasing the same claim: that an AI system can be trusted to finish a task rather than merely draft it.

Practical implications

Early customers named at launch include Zapier, RingCentral, Virgin Atlantic, and Nvidia. Finance teams at these companies are using the tools to speed up month-end reporting and forecasting by automating the work of pulling data, reconciling figures, and building presentations. For AI builders, this pattern is worth studying: the most immediate enterprise use cases for agents are not creative tasks but structured, repeatable workflows in finance, sales, and operations.

OpenAI says nearly all of its own staff now use the agentic products daily. Sam Altman has repeatedly reset usage caps as a nudge to push adoption higher. If you are building agent-based products, expect OpenAI to continue iterating on pricing and controls to capture enterprise budgets.

Caveats

The 10 million figure is a self-reported company estimate, not an audited metric. It counts weekly active users across Codex and ChatGPT Work combined, not paying seats, enterprise licenses, or the number of agents actually run. OpenAI's usage claims have drawn skepticism before. The figure also does not reveal how many of the 10 million sit on paid plans or how often they return once the novelty of a new model fades. For now, the number reads as a marker of momentum rather than proof of a market OpenAI has already won.