We use cookies to keep you signed in. Cookie policy
We use cookies
CalculatorAI uses cookies to keep you signed in and remember your preferences. Optional cookies help us understand how the site is used. Cookie policy
How Many Tokens Is Your Text?
Total tokens
Paste some text to count it
Nothing to count yet
Paste a prompt or a reply above, upload a screenshot, or enter exact token counts in Advanced settings — then this fills with the token count and what it costs.
Drop in a paystub, bank statement or screenshot — the AI reads it and fills these fields for you.
Income & Expense Tracker is where this stops being a one-off number: keep your real balances there and the projection updates itself.
Describe an invoice and the AI writes it. PDF without watermarks, free.
Most token calculators multiply one count by one price. That is wrong twice over: the count changes with the model, and the cheapest rate card is not always the cheapest request. This page counts per model and sorts by what your text actually costs. Rates come from the providers' own pages, OpenAI and Anthropic among them.
Prompt shapes differ wildly. Paste the thing you actually send — a long system prompt behaves nothing like a one-line question when it meets a different tokenizer.
Output costs two to six times what input does everywhere. If your replies are long, the output rate decides your bill and the input rate is almost decoration.
Outside OpenAI, no tokenizer runs in a browser. A single confident number for a Claude or Gemini prompt is invented precision — budget against the top of the range.
Switching between vendors, or between Claude generations, changes how the same words are split. A 30% swing in count can wipe out a 30% saving in price.
Count on the model you will actually use, price it on your own text, and read the range rather than the midpoint.