About CostUse
CostUse turns published usage prices for AI models, APIs, and other metered services into workload-based cost estimates. Use the calculators to compare a defined workload before choosing a service or setting a budget.
AI model pages include token-volume tables and an optional text estimate. The text estimate uses roughly four characters per token. It is a planning shortcut, not a provider's official tokenizer.
What we publish
- AI model cost tables based on provider-published input and output token rates
- Calculators for APIs and other metered services when prices can be expressed as public usage units
- Workload estimates that show the rate, billing unit, and assumptions behind the calculation
Rates, sources, and estimates
Each pricing entry links to a provider's public pricing or documentation page and records a last-checked date. Rates can change after that check, so the provider's current page and your contract determine what you will pay.
An estimate uses the rate and workload assumptions shown on its page. It may not include caching, batch rates, free credits, contract discounts, regional terms, taxes, or additional fees.
Independence
CostUse is an independent reference site, not an official billing or pricing service for the providers it covers. A provider's inclusion does not mean that it endorses CostUse or sponsors an estimate.
Who maintains it
CostUse is maintained by Kevin Henrique. Send a pricing correction or calculator suggestion through the contact page.
Frequently asked questions
Are these official invoices?
No. They are planning estimates based on public list prices.
Why can my bill differ from an estimate?
A provider may use different tokenization, model versions, regional rates, caching, batch pricing, free credits, contract discounts, taxes, or fees. A simple calculator may not include those conditions.
Can I request a calculator?
Yes. Send the service name and its public pricing page through the contact page.