Your goal: compare a workload
- Select models in Compare and choose Calculate cost.
- Select one complete endpoint per model; dashboard minimum input/output prices can belong to different providers.
- Enter each model's own native counts and number of identical requests.
- Use Add request profile when lengths differ. Compare complete profile totals.
One agent task can make many API requests. Do not enter an entire log's summed input as one request length: it may falsely exceed a context limit and hide per-request pricing conditions.
When input is cached
Total input includes uncached input, reads and writes. Specify actual cache counts; unchecked means all supplied input is uncached in this profile. Read/write categories cannot overlap. Write and other unreconciled adapters remain gated.
Equal token budgets compare rates, not equal text or task performance. See provider routing and cache accounting.
Sources and further reading
These explanations describe Examenos's data and display policy. Calculator examples are explicitly hypothetical. Inspect dated endpoint evidence and the formulas; projections do not authenticate a bill.