OpenAI adds a credit-usage console and self-service limit requests to ChatGPT Enterprise

What the Global Admin Console now shows

The core of this release is a single screen. OpenAI is folding ChatGPT and Codex credit consumption into one view inside the Global Admin Console, with a billing tab and an overall plan view. Instead of reasoning about spend in the abstract, admins get a breakdown across users, products, and models.

The described capabilities are concrete: track usage and credit trends over time, identify top users and emerging patterns, and break spend down by user, product, and model. For teams that want to leave the dashboard behind, the same data is exposed through a unified Cost API for analysis in their own systems.

The framing OpenAI uses is telling — the analytics are meant to help admins distinguish 'increased usage driven by valuable work' from 'usage patterns that may require closer review.' That is a diagnostic tool, not just an invoice.

The spend-control model: default, group, override

OpenAI had already shipped granular credit usage limits for custom roles earlier in 2026. This update layers a hierarchy on top: a default limit for the whole workspace, configurable limits for specific groups, and individual overrides for people who need more capacity.

That three-tier structure is designed to solve a specific problem the announcement names directly — letting individual power users keep working without raising the ceiling for everyone else. It is a policy engine expressed through the admin UI rather than a blunt per-seat cap.

The end-user request flow is the notable frontend piece

Most of the coverage-worthy frontend work is on the employee side. Users in these workspaces can now open their workspace settings and see their credit usage against their available budget. When they hit a wall, they can request additional credits and attach context about what they're working on.

That context field turns an approval into an informed decision rather than a rubber stamp. It also changes the interaction pattern: the friction of asking for more is moved into the product itself, so a power user's request and an admin's judgment happen in the same system that generated the spend.

We asked the team at OpenAI to build usage analytics to help find and train-up folks who haven't adopted Codex, and for granular usage controls to keep spend predictable. These new tools are helping us faster scale productivity of our employees while keeping safeguards in place. —Ryan Oksenhorn, Co-Founder, ZiplineMontana Labs

The Zipline quote reveals the dual purpose behind the analytics. Oksenhorn describes using them not only to cap spend but to locate employees who haven't adopted Codex and train them up — adoption tracking and cost control drawn from the same consumption data.

What this changes for teams deploying ChatGPT Enterprise

Credit consumption is now a first-class, observable dimension of an enterprise deployment, with both an admin dashboard and an API behind it. The specific implication for applied teams is that budget governance no longer has to live in a separate finance spreadsheet reconciled after the fact — usage attribution by user, product, and model, plus an in-product request-and-approve loop, is available in the workspace itself starting today.