Free tool
LLM Cost Calculator
Estimate your monthly AI spend from requests, tokens and price per million tokens, including prompt caching and retries.
Calculated in your browser. Nothing you enter is sent or stored. Fields marked * are needed.
3. Estimate, 30-day month
Still needed: input price, output price, requests per day, input tokens per request, output tokens per request.
How the estimate works
Monthly requests are requests per day times 30, plus the share sent again as retries. Input cost is monthly requests times input tokens per request times the input price per million tokens, with the cached share billed at the cached-input price instead. Output cost is monthly requests times output tokens per request times the output price. The retry line shows how much of the total comes from requests sent more than once.
Everything is calculated in your browser from the numbers you enter. Nothing is sent or stored. Preset prices are standard list prices read from each provider's pricing page on 18 November 2026; your contract, batch or regional prices may differ.
Questions
What is a token? The unit a model reads and writes. Providers bill input tokens (what you send, including the system prompt and any retrieved context) and output tokens (what the model returns) at different prices, usually per million tokens.
Why are retries a separate line? When a request times out or fails and is sent again, the provider bills both attempts. If several teams share a gateway, make sure each retry is charged to the team that can prevent it.
What does prompt caching change? Input that repeats across requests, such as a long system prompt, can be billed at a lower cached-input price where the provider supports it. The calculator applies that price to the share of input you enter as cached.
How well does your organization manage AI spend?
Take the free AI Economics Signal, a 15-minute self-assessment of how your organization manages AI spend.