Apps reuse model outputs to run faster

When an app runs, it automatically reuses model outputs already computed by another app in the same org instead of re-running them. Compute costs drop and run times improve, with no setup required.