Cursor raises team limits and introduces Premium tier for power users

Cursor, the developer of the popular AI programming environment, announced an increase in usage limits for the Teams plan and the launch of a new Premium plan designed for users with the highest AI agent workloads. The changes take effect immediately for all new customers and for existing customers starting with billing cycles that begin on July 1, 2026.

What Changes in the Teams Plans

The main change is the introduction of a separate usage pool for the Composer model, Cursor's latest development. According to the company, Composer 2.5 delivers frontier-level performance at significantly lower cost compared to previous flagship models. Each seat in the Teams plan now receives two separate usage pools: one for Cursor's own models and Auto mode, the second for Third-Party API—models from other developers that can be connected through Cursor. According to the company, this increases the available usage volume for a standard seat without changing its price.

How the New Premium Plan Works

According to a recent Cursor Developer Habits Report, a small share of "power users" in any team accounts for the majority of spending and unpredictable overage costs. To reduce these costs and increase budget predictability for such users, the company introduces a Premium seat: it provides five times (5x) more included usage compared to Standard, but costs only three times (3x) more. Teams can freely mix seat types. According to Cursor's estimate, the Composer pool in the Premium plan should be sufficient for a full month of intensive agent work for 99% of users.

Updated Teams plan conditions:

• Standard: increased total usage volume; $32 per seat per month with annual billing, $40 with monthly billing.
• Premium: 5 times more usage than Standard; $96 per seat per month with annual billing, $120 with monthly billing.
• Changes take effect for new customers immediately, for existing customers starting with billing cycles from July 1, 2026.
• Key model for calculation — Composer 2.5.
• Data source for user behavior — Cursor Developer Habits Report.

Why New Spending Transparency is Needed

Beyond changing the plans themselves, Cursor updated the dashboard, which now shows users how close they are to their limits—separately for the Auto + Composer pool and for third-party API models. The system will also recommend to administrators which seat type best suits a particular employee based on their actual usage patterns. Separately, the company reworked its spending alert system: administrators can now set "smart" thresholds in dollars and receive notifications via Slack or email before the billing period invoice becomes a surprise.

Taken together, these changes reflect a broader problem faced by all providers of agent AI tools for development: as agents take on more autonomous work—not just code autocomplete, but full task execution—inference costs become significantly less predictable than with a classic subscription model. Separate pools, premium seats, and active spending notifications are Cursor's attempt to balance flexibility for heavy users with cost control for teams and their finance departments.

Cursor's changes in pricing reflect the maturation of the market for AI development tools overall: if in the early stage such products competed primarily on functionality and code completion quality, by 2026, when agents take on more independent work—from writing entire modules to multi-step debugging—the key competitive field becomes precisely the management of inference costs at the model level. Companies developing such programming environments are forced to balance between an attractive price for the mass user and covering real computational costs for the small share of customers who use agents at the limits of capability almost continuously.