# See what every agent costs, *and bring it down.*

Slide 6 of 10 · Launch is day one

Spend by agent, model and person. Prove a cheaper model still passes before you switch.

![Monitor, Cost tab in Mindset: total spend, calls, average cost per call and spend over time](https://docs.mindset.ai/images/decks/launch-is-day-one/monitor-cost.png)

## Speaker notes

Every model call is recorded and priced, and every breakdown adds up to the same total. The biggest savings, in order: prove a cheaper model still passes your tests; turn fixed work like arithmetic into a function; return less from each system call; schedule for the rate the data actually changes. There is no spending cap to set in Mindset; spend is bounded by run limits.
