Cursor splits usage into Cursor Models and Other Models. The Cursor Models pool includes Grok 4.6, Grok 4.5, and Composer 2.5. Other Models covers third-party options such as Claude and GPT inside Cursor. India’s Start plan at ₹649/month tax inclusive includes only Cursor Models, non-fast, with Grok effort fixed at medium. Pro, Pro+, and Ultra include both pools. Teams pay a $0.25 per million token Cursor Token Rate on third-party model requests; Cursor’s own models are exempt.
Why pools exist
Pools let Cursor price its own models differently from third-party API models. Start customers stay inside Cursor Models for predictable INR pricing. Paid USD plans open Other Models when you need those vendors’ strengths.
Switching pools mid-task only works if your plan allows the target model. Teach the team the pool names so nobody files tickets claiming Start is broken when Claude is simply excluded. Route exploratory coding to Cursor Models when quality is fine, and reserve Other Models for harder reasoning tasks.
Start constraints inside Cursor Models
On Start you cannot enable Fast mode or change Grok effort away from medium. That is a hard product limit, not a settings bug. If you need Fast or variable effort, budget for a higher plan.
Composer 2.5 remains available in the Cursor Models pool on Start for multi-file drafting within those rules. Route exploratory coding to Cursor Models when quality is fine, and reserve Other Models for harder reasoning tasks. On Teams, forecast the $0.25 per million token rate whenever Other Models usage climbs month over month.
Teams and token rate
- Third-party requests: API price + $0.25/MTok Cursor Token Rate.
- Cursor’s own models: exempt from that rate.
- Finance should forecast Other Models spend separately.
On Teams, forecast the $0.25 per million token rate whenever Other Models usage climbs month over month. Teach the team the pool names so nobody files tickets claiming Start is broken when Claude is simply excluded. Route exploratory coding to Cursor Models when quality is fine, and reserve Other Models for harder reasoning tasks. Write the intended model pool and plan name into your notes before the session so you do not discover a Start exclusion mid-task. Keep git status clean before agent runs, and recheck Cursor pricing when your seat or model needs change.
Related on this site
Frequently asked questions
-
Is Composer 2.5 an Other Model?
No. Composer 2.5 is listed in the Cursor Models pool along with Grok 4.6 and Grok 4.5. That is why it appears on India Start. Name the pool in standup notes so teammates stop confusing plan bugs with model exclusions.
-
Can I buy Other Models on Start?
Start does not include the Other Models pool. Upgrade to Pro or higher when you need Claude, GPT, or other third-party models inside Cursor. Name the pool in standup notes so teammates stop confusing plan bugs with model exclusions. Verify the live Cursor account meter before you promise capacity.
-
Do Tab completions use pools the same way?
Tab is called out as unlimited on paid plans. Model-backed chat and agent calls still respect which pools your plan includes. Check usage UI for how your seat meters requests. Name the pool in standup notes so teammates stop confusing plan bugs with model exclusions.
-
Does the token rate apply to Grok?
Cursoru2019s own models are exempt from the $0.25/MTok Cursor Token Rate. That rate applies to third-party model requests on Teams atop API pricing. Name the pool in standup notes so teammates stop confusing plan bugs with model exclusions. Verify the live Cursor account meter before you promise capacity.
Sources
Bdeb Technology builds websites, WordPress systems, and tools on top of models like these. A written quote comes back within 24 hours.