ChatGPT Enterprise gets spend controls
OpenAI added usage analytics and per-team spend limits to ChatGPT Enterprise, so admins can see who burns credits and cap it without one blanket limit.
OpenAI on June 18 added credit usage analytics and new spend controls to ChatGPT Enterprise, and both are available to admins today. The change is aimed squarely at the question every team asks once the trial ends: where is the money actually going.
The Global Admin Console now pulls ChatGPT and Codex credit usage into one view, broken down by user, product, and model. Admins can track trends over time, spot the heaviest users, and pull the same numbers through a Cost API to chew on in their own systems. On the control side, you can set a default credit limit for the whole workspace, override it for specific groups, and override it again for individuals who genuinely need more. Employees see their own usage against their budget and can request a bump with a note explaining why.
What it means for operators
This is the unglamorous plumbing that decides whether an AI rollout survives its second invoice. The useful shift is the move away from a single blanket cap: one power user on Codex no longer forces you to raise the ceiling for the entire company.
If you are sizing up tools, predictable spend belongs on the checklist next to capability, the same way you would weigh any AI tool for a small team or build out a solo operator stack. It also lands in a month where pricing models keep shifting under buyers, as when Anthropic paused token billing for its Agent SDK.