Metered pricing
Paying per run rather than per month, usually because the maker's own cost scales with use.
It is the right model when cost genuinely scales with use, and it is the easiest model to be surprised by. Where a listing carries a measured cost per run it comes from a real sandbox execution, which is why it is often higher than a maker's documentation suggests; no listing has been run there yet.
The number that matters is cost per run multiplied by realistic monthly volume, including retries. A task set with a fifty seven percent success rate runs nearly twice.
Unbounded inference cost is the most common way a first listing produces a refund and a bad review at the same time.
Estimating volume from what you plan to do rather than from the thing the agent replaces.
Related terms
See Metered pricing on a real listing
Every term here shows up in the catalogue next to a real result, with the findings published and the limits stated. Free to browse, no account needed.
Open the catalogue