Tokenizer Visualizer
Large language models don't read words — they read tokens (chunks of characters). Type below to see how your text gets split up. (This is an approximation of GPT-style tokenization.)
Tokens
≈ a rough rule of thumb: 1 token ≈ 4 characters ≈ ¾ of a word in English. Models bill and "remember" by tokens, not words — which is why long prompts cost more.
The energy figures are rough educational estimates — real usage depends heavily on the model, the hardware, and the data center. Edit the two values above to match any source you trust.