Token and cost estimator
Paste text below. Get an estimated token count and what it costs to process.
How this estimates
Chinese characters → 1 token each (conservative) CJK punctuation → 1 token each Other characters → 1 token per 4 characters Then output tokens are added per the task you pick
This is a heuristic, not a real tokeniser. Expect ±10% against the API's own count.
For exact figures, send the text to the API and read usage.prompt_tokens.