Ultimate Plan Question

Support Help
1

Hi Guys,
Right now im on the Ultimate Plan, used 100% of it
also topped out 250$ On-Demand Usage
I’m using a lot of cloud agents + Automations
there is a way to upgrade to a better plan?
Im Individual user under my own company.
Thanks

5

Hey Netanel,

Quick clarification: the plan you’re on (what you’re calling “Ultimate”) is Ultra, and it’s currently our highest individual plan - there isn’t a tier above it right now. So the way to keep going past your included usage is on-demand, not a bigger plan.

A few options:

  1. Raise your on-demand spend limit. The $250 is a cap, not a hard ceiling - you can raise it or set it to “No Limit” in the Spending tab of your dashboard. On-demand is billed at the model’s API rate with no markup. More detail: Spend limits and Usage and limits.
  2. Lean on Cursor’s own models - Cursor Grok 4.5 and Composer 2.5 draw from a separate included pool tracked apart from third-party models, so using them for routine agent/automation work stretches your budget much further.
  3. If you’d like an actual higher tier, there’s an open thread where others have asked for exactly that - feel free to add your voice: More Usage may be a plan higher than Ultra. We’re tracking demand there.

And if you’re scaling this up under your company and want pooled or custom usage, our team can walk through Teams/Enterprise options: contact sales.

6

Hi,
I’m using only cursor grok 4.5 and composer 2.5
What takes a lot of from my budget is bugbot, pr approval, and the security agents.
There is no option to manual select the model they use.
All other of my custom automations are set to grok and composer

8

Ah, that’s the difference: for the managed agents (Bugbot, PR approval, security review) the model is set on our side, so there’s no picker like on your custom automations. You’re not missing a setting.

To cut their spend, control how hard/often they run instead of which model:

  • Bugbot: set effort level → Default (or Custom to only go deep on risky changes), turn on Incremental Review (reviews only new changes, not the whole diff each push), and use run only when mentioned (cursor review / bugbot run) if you don’t need every push reviewed. All in your Bugbot settings.
  • PR approval + security: no model picker either - the lever is running them selectively (tighten their triggers) rather than on every change.

Picking the model yourself is a fair ask - we’re tracking interest in it. No timeline, but I’ll follow up here if that changes.