Cursor Teams Pricing Overhaul: Standard $32 vs Premium $96 — Which Seat Saves You Money?

Cursor Splits Teams Into Two Tiers

Cursor has restructured its Teams plan effective June 2026. Instead of a single per-seat price, companies now choose between Standard seats at $32/month (annual) or $40/month (monthly) and Premium seats at $96/month (annual) or $120/month (monthly). The change is live immediately for new customers and kicks in July 1 for existing renewals.

The logic behind the split is straightforward: Cursor's own "Developer Habits Report" showed that power users — roughly the top 20% — drive the majority of compute spend. Rather than raising prices across the board, Cursor is letting teams assign seat types based on actual usage patterns. A team of 10 might run 7 Standard seats and 3 Premium seats instead of paying a uniform rate that overcharges light users and underserves heavy ones.

Two Separate Usage Pools

Each seat type includes two distinct usage pools. The first is the Composer/Auto pool, which covers Cursor's first-party models including Composer 2.5. The second is the Third-Party API pool, which covers calls routed to Claude, GPT, Gemini, and other external models. Premium seats get approximately 5x the allocation in both pools compared to Standard.

Cursor also introduced a real-time spend dashboard with Slack and email alerts when usage approaches pool limits. This is a direct response to the "bill shock" complaints that plagued usage-based pricing at other AI tools. Teams can now see exactly who is consuming what, in real time, before overages hit.

Composer 2.5: The First-Party Cost Advantage

The pricing restructure coincides with Cursor pushing its own Composer 2.5 model as the default coding engine. Cursor claims it delivers "frontier performance at a fraction of the cost" — and the numbers back this up. Composer 2.5 Fast runs at $3.00/$15.00 per million input/output tokens, while Composer 2.5 Standard runs at $0.50/$2.50. Compare that to Claude Sonnet 4.6 at $3.00/$15.00 or GPT-5.4 at $2.50/$15.00.

The strategic play is clear: Cursor wants teams using Composer 2.5 Standard for the bulk of their work at one-sixth the cost of frontier third-party models. Premium seats make economic sense primarily when developers need heavy access to Claude Opus 4.8 ($5.00/$25.00) or GPT-5.5 ($5.00/$30.00) through Cursor's third-party pool.

Direct API Cost Comparison

Let's put concrete numbers on what each seat buys versus going direct to APIs. A Standard seat at $32/month — if we assume roughly half goes to the Composer pool and half to third-party — gives you about $16 worth of third-party API credits. At Claude Sonnet 4.6 rates ($3/$15 per 1M tokens), that's roughly 5.3M input tokens or 1.07M output tokens per month.

Premium gives you 5x the usage at 3x the cost — a clear discount per token. But it only pays for itself if you're actually consuming that volume. A developer generating under 1M output tokens/month on third-party models is overpaying at $96.

The Crossover Point

The math simplifies to this: Premium pays for itself when a developer consistently uses more than ~2M output tokens per month on third-party models. Below that, Standard plus occasional direct API calls is cheaper. Above it, Premium's bundled rate beats pay-as-you-go by 20-40%.

For context, Cursor's own data shows power developers generate 8,600+ lines of code per week. At ~10 tokens per line plus 4x context overhead, that's roughly 3.4M total tokens per week — well into Premium territory. But developers who primarily use Composer 2.5 Standard for routine coding and only reach for Opus or GPT-5.5 occasionally will save with the Standard seat.

What This Means for Team Budgeting

The two-tier approach is better than uniform pricing, but it creates a new management overhead: someone has to decide who gets Premium. The spend dashboard helps, but expect friction in the first quarter as teams figure out their actual distribution. The smart play is starting everyone on Standard and upgrading based on 30 days of real usage data — the dashboard makes this possible without guessing.

For teams evaluating Cursor against alternatives, the effective per-developer cost now ranges from $32 to $96/month — a wide band. Compare that to GitHub Copilot's token-based billing or Claude Code's direct API pricing at $3/$15 for Sonnet. The right answer depends entirely on how much your heaviest users actually consume.