Hi @DRobb, and thanks for the nudge, everyone.
There’s no separate VM, compute, or memory charge for Cloud Agents. Usage is billed at the selected model’s API rate, the same rates listed on our models and pricing page, drawing from your plan’s included usage first and then continuing on on-demand usage at those same rates. Builds, which prepare the environment agents start from, are included at no additional cost.
Two things that do affect cost: the model and context window you choose, since both change how many tokens a run consumes, and on Teams and Enterprise plans the Cursor Token Rate on third-party models, which applies to all usage on those plans rather than to Cloud Agents specifically.
On whether this is temporary, I can only speak to what’s published today, which is the structure above.
https://cursor.com/docs/cloud-agent