Cursor shipped a new set of Enterprise admin tools on April 26, adding model controls, updated spend management, and more granular usage analytics. The update targets a growing pain point: companies adopting AI coding tools at scale need visibility into what models their teams use and what it costs.
Model controls let Enterprise admins decide which AI models are available to their developers. This matters because model choice directly affects both output quality and per-seat cost — a team defaulting to the most expensive model for every autocomplete burns budget fast. Admins can now set guardrails instead of relying on individual developers to pick cost-appropriate models.
TIP
If you’re an Enterprise admin, audit which models your team actually uses before setting restrictions. Usage analytics from this same release gives you that data — set controls based on real patterns, not assumptions.
The spend management update gives finance and engineering leads better tools to cap and track AI spending across their organization. As companies scale from a handful of Cursor seats to hundreds, unmanaged AI tool spend becomes a real line-item problem. Structured spend controls — policies and tooling that let teams monitor, limit, and approve spending — are standard practice in procurement, and Cursor is now building that expectation into its Enterprise tier.
Usage analytics got more detail, though Cursor’s changelog doesn’t specify exactly which new metrics are available. The direction is clear: admins need to see per-team or per-user consumption patterns to justify renewals, right-size seat counts, and identify underutilized licenses.
The practical takeaway for Enterprise Cursor customers: log into your admin dashboard and review the new controls now. Model restrictions and spend caps are the kind of settings that save money retroactively — every week without them is a week of unmanaged cost. If you’re evaluating Cursor for a larger team, these admin features close a gap that previously made it harder to justify against more traditional IDE tooling with established governance features.