CloudCostCompareEsta página ainda não está disponível em Português (Brasil), então o texto abaixo está em inglês. Abrir a página em inglês
AI model costs
Language models are priced per token, which is hard to picture. These tools turn a task you recognise, such as a support chat or a document summary, into a monthly cost on each model, using the rates the vendors publish.
Task cost estimator
Pick a task, such as summarizing documents or running support chats, set a monthly volume and see what each model costs. The formula is shown for every row.
Published discounts
List prices next to the prices with prompt caching and batch processing, using only the discounts each vendor publishes.
What a token is
A model reads and writes text in pieces called tokens. In English a token is about three quarters of a word, so a page of 500 words is roughly 670 tokens. Vendors charge separately for the tokens you send in (input) and the tokens the model writes back (output), per million tokens. Output costs more per token because the model produces it one token at a time.
Two things catch people out. A chat sends the whole conversation again on every turn, so the input grows with each message. And some models spend extra "reasoning" tokens before answering, which are billed as output. The estimator models both.