Cursor enhances Teams plan with predictable pricing and new Premium seat options

Cursor's June 2026 Teams pricing update split usage into separate first-party and third-party pools and added a Premium seat for power users. The headline was cost clarity: admins could finally see who was burning through what, and heavy users became a bounded cost instead of a billing surprise. If you run LLM agents at production scale, that structure is worth understanding — even though the exact dollar figures have since moved on, so treat the numbers below as the June 2026 launch terms and check Cursor's current pricing page before you budget.

What changed in the June 2026 Teams update

The update made three surgical changes aimed at predictability, all corroborated by the StartupHub report on the announcement, which cites Cursor's own announcement post:

• Pool split: usage for Cursor's own models ("Composer and Auto") is tracked separately from third-party API usage. Before, everything metered from a single pool — now first-party and third-party consumption are independently visible.
• Premium seat: a new tier offering five times the included usage of Standard at three times the price, aimed at heavy users and meant to cover roughly a full month of intensive agent use.
• More usage for Standard: Standard seats got increased total usage allowances at the existing price point.

Timing-wise, the changes applied immediately for new customers and for existing Teams orgs at their first renewal on billing cycles starting July 1, 2026.

How the pool split works in practice

Each Teams seat now carries two separate allowances: one for "Composer and Auto" (Cursor's first-party models) and one for "Third-Party API" usage (requests routed to external models). Consumption in one pool never eats into the other.

Consider a typical split team:

• Engineering leans on Composer for boilerplate and refactors — that burns only the first-party pool.
• Another group prefers frontier third-party models for tricky reasoning tasks — that burns only the third-party pool.
• Before the split, a spike in third-party usage could throttle everyone's core workflows. After it, overruns stay localized to the pool that caused them.

For admins, the practical wins are independent tracking per pool, forecasts that stay granular (actuals-vs-budget per pool), and the ability to tune seat allocation in one pool without disturbing the other. The StartupHub report adds that admins get real-time usage visibility in the Cursor dashboard, with progress toward limits for both pools, plus rebuilt spend alerting with dollar-based thresholds and Slack or email notifications.

The Premium seat, and who it is for

The Premium tier exists to take cost ambiguity out of power-user workflows. At launch, the terms were:

• Usage: 5× the included usage of a Standard seat.
• Price: $96 per seat per month on annual billing ($120 monthly) — exactly triple the Standard seat's $32 annual / $40 monthly at the time.
• Pitch: lock in enough usage for your busiest engineers instead of paying unpredictable overages.

The intended user is the heavy contributor: long-context completions, agent chains, high-iteration prototyping — anyone who has blown past their allowance more than twice in a quarter. The math that matters is the ratio (5× usage for 3× price), which favors heavy, concentrated use over simply buying more Standard seats.

One caution: those dollar figures are June 2026 launch terms. Cursor's live pricing page now lists Teams at $40 per user per month with Standard and Premium tiers, centralized billing, usage analytics, SSO, and team marketplace features. Pricing has moved since launch, so re-verify current numbers before forecasting — the pool structure and the premium-ratio logic are the durable parts.

Where Composer 2.5 fits in

The pricing update shipped alongside Composer 2.5, positioned as frontier-quality model performance at reduced cost. Because first-party usage is now capped and tracked in its own pool, model efficiency gains land directly in that pool: every unit of allowance buys more output. Both Standard and Premium seats got access, so the improvement effectively rewards teams that lean on Composer rather than defaulting to external LLMs for every task. For Standard users hovering near historical usage ceilings, that efficiency gain is a quiet second price cut.

How to apply this on your team today

No CLI wizardry needed — the controls live in the dashboard:

1. Audit your pools separately. Check how much of your burn is first-party versus third-party. If third-party dominates, that is a prompt-engineering or model-routing problem, not a seat problem.
2. Identify your 5× users. Anyone consistently exceeding Standard allowance is a Premium candidate — compare triple seat cost against their trailing overage spend.
3. Set spend alerts. Use the dollar-based thresholds with Slack or email notifications so finance hears about spikes before the invoice does.
4. Route routine work to first-party models. Composer-class models on boilerplate stretch the cheaper pool; reserve third-party calls for tasks that genuinely need them.

Teams that ship AI-assisted code to production should treat seat planning like capacity planning — the same discipline you apply to background jobs and deploy pipelines. If your shipping process is solid but your AI tooling spend is a black box, start with our production background jobs guide and the ship-to-production checklist to put the same rigor around cost that you already put around uptime. And if Row Level Security is part of your stack, the Supabase RLS production checklist pairs well with per-pool usage reviews.

Bottom line

Cursor's 2026 Teams update did what most AI SaaS pricing changes only claim: it put predictability into the hands of admins and finance leads. Separate first-party and third-party pools, bigger Standard allowances, a bounded-cost Premium tier for heavy users, and dashboard-level spend controls turned AI tooling from a budget unknown into a forecastable line item. The exact prices have evolved since launch — always confirm against cursor.com/pricing — but the structure is the real story: meter separately, cap the heavy users, alert before the invoice.

Ship AI features your team can afford to run. OTF kits give your coding agent production-ready templates with real architecture — browse the kits.

Sources

• Cursor pricing (official, current terms)
• Cursor Teams upgrades pricing for predictability — StartupHub
• Cursor Teams pricing announcement, June 2026