A routed token is any input or output token from a request covered by an AI Router routing decision, regardless of who executes the request. This means the same usage measure applies whether AI Router proxies the request or your application calls the selected model directly. It also applies when your application runs a different model than the one selected: the tokens are still routed tokens.
Decisions made in shadow mode are the exception. A shadow decision is an observation, not a routing decision, so its tokens are not routed tokens: they are not counted toward your quota, not estimated when no report arrives, and not counted as savings.
Plans include a monthly routed-token quota; usage beyond the quota is billed as overage rather than cut off, and routing is never throttled. Model costs are not part of this: with BYOK they remain on your own provider bill, the AI Router plan price covers the routing fee.
Usage comes from three sources:
Billing is based on measured usage. Where a routing decision (not a shadow decision) has no usage report, we estimate its usage conservatively from your measured usage profile; the basis of that estimate is fixed in your contract. Reporting actual usage gives you the most accurate record and avoids estimates altogether.
Usage reports support more than accurate billing. Savings and opportunity reports compare routing decisions with the requests your application actually executed, so those reports are only complete when selection usage is reported back for every routing decision.
Only usage that ran on the model the router selected counts as savings. A report whose model or reasoning effort differs from the selection is a deviation: its tokens stay routed tokens, but they are not counted as savings of that routing decision. See how reports are matched to decisions.
For shadow decisions, usage reports are the baseline of the pilot report, which compares what your application ran with what the router would have chosen.
Open your team in the dashboard and go to Usage. It shows the current month's routed tokens split into measured, reported and estimated usage, your quota, and a month-by-month history.
See the pricing page for plan details or the pricing FAQ for short answers to common usage and billing questions.