Token Lab: what an AI answer costs

Every AI answer is paid for in tokens. Pick a model, ask it something, and see the answer beside its bill. Then ask the same thing of another model and compare.

How it's priced

  1. Text is counted in tokens. A token is a piece of a word - about four tokens for every three English words. "Photosynthesis" is a few tokens; "the" is one.
  2. What you send and what comes back are priced separately. Your prompt, plus any instructions the activity adds, is input. The answer is output, and output usually costs several times more per token, because the model writes it one token at a time.
  3. Each model has its own prices. Bigger, more capable models cost more per token - often ten or a hundred times more than a small one. A longer answer costs more on any model.

The models you can try

Loading the price list…

Try it

Your prompt goes to the company that makes the model you pick, to be answered, and nowhere else. Nothing is saved. Please don't include anyone's name or other personal details - prompts with personal information are declined.

Ask anything - try a short question, then ask for the same thing in more detail.

0 / 600

Things to try

  • Ask a short question, then ask for "a detailed explanation with examples". Watch the output side grow.
  • Ask the same question of the cheapest and the dearest model. Is the dearer answer worth the difference?
  • Paste a paragraph and ask for a one-sentence summary. A long prompt with a short answer is cheap on output.
  • Ask for a list of ten items, then for three. The answer's length is the part you control most.