LLM Cost Calculator

How it works

1. 1

   Pick the model

   Select the model you call from the dropdown. The per-1M input and output rates update underneath so you can see what you are paying for before you enter a single token.

2. 2

   Enter tokens per call

   Set the input (prompt) and output (completion) tokens for a typical request. Drag the sliders for quick what-ifs, or type exact counts from a usage log. If the model supports prompt caching, add the cached share to price it at the lower rate.

3. 3

   Scale to your volume

   Enter how many calls you expect — a batch job, a daily total, or a month of traffic. The hero figure shows the combined cost; the stat grid keeps the per-call number and the input/output split in view.

Frequently asked questions

How does this AI API cost estimator calculate the price?

It multiplies your input tokens by the model's input rate and your output tokens by the output rate, both quoted per million tokens, then sums them for one call and multiplies by your call count. Cached input tokens use the provider's cheaper cached rate when one exists.

Why are input and output priced differently?

Providers charge more for generated (output) tokens than for the prompt (input) tokens they read. GPT-4o, for example, bills output at four times the input rate, so a chatty completion costs far more than a long prompt with a short answer.

How do I get the token counts to enter?

Use your provider's usage dashboard or the token count returned in each API response. This token cost calculator does not tokenize text itself — it takes the counts you supply and does the pricing math, which keeps it accurate for any model.

Are the prices current?

Rates come from a shared table dated on the page under the result. Providers change pricing often, so treat the number as a close estimate and confirm against the official pricing page before you commit a budget.