Prompt Token Estimator
See a prompt's token size and cost before you send it.
It helps to know what a call costs before you make it. This prompt size calculator lets you estimate tokens before an API call: paste the prompt you plan to send, pick a model, and read a live token count with the input cost worked out from the model's rate. Add a guess for the reply length and it prices the output too, so you get a cost per call rather than just a token number. The estimate uses a character-and-word heuristic, not the model's exact tokenizer, so treat it as a close planning figure.
Your prompt
Your prompt never leaves this page. The estimate is worked out in your browser and nothing is uploaded.
Estimated prompt tokens
49
Approximate — real tokenizers differ by model
- Input cost
- $0.000122
- Output · 150 tok
- $0.001500
- Cost per call
- $0.001622
Prompt size
- Characters: 199
- Words: 36
- Blended token estimate: 49
Prices updated January 2026. GPT-4o bills $2.5/1M input and $10/1M output. Verify before you rely on a quote.
This estimates tokens before you call the API, so you can size a prompt and its cost up front. The count is a heuristic, not an exact tokenizer, and output length is your own guess — treat the cost as a planning figure, then confirm with the token counts your API response returns.
How it works
- 1
Paste your prompt
Drop in the exact text you plan to send — instructions, examples, the question. The token estimate updates on every keystroke, so you can trim wording and watch the number fall.
- 2
Pick the model and reply length
Choose the model you'll call, then set roughly how many output tokens you expect back. Leave output at zero to price the prompt on its own.
- 3
Read the cost per call
The headline is the estimated prompt tokens. Underneath you get the input cost, the output cost for your expected reply, and the combined cost of one call.
Instant & 100% private — nothing is uploaded
Every calculation runs locally in your browser. The prompts, token counts and numbers you enter stay on your own device and are never sent to a server — nothing is stored, logged or shared.
Frequently asked questions
- How do I estimate tokens before an API call?
- Paste the prompt here before you send it. The tool counts characters and words and blends two rules of thumb — about four characters per token and 1.33 tokens per word — into a single estimate, then prices it against the model you pick. It runs entirely in your browser, so you can check a prompt without spending anything.
- How accurate is the token estimate?
- For ordinary English it's usually within 10–20% of the model's real tokenizer. It can't be exact because each model splits text with its own vocabulary, and code, non-English scripts and heavy punctuation tokenize differently. For the exact figure, read the token counts your API response returns.
- Why does it ask for an expected reply length?
- Output tokens are billed separately, almost always at a higher rate than input, so the full cost of a call is your prompt plus the model's answer. Since you can't know the reply length in advance, you enter a rough guess to get a realistic cost per call rather than input alone.
- Does my prompt get uploaded?
- No. The counting and pricing happen in your browser with JavaScript. Nothing you paste is sent to a server, logged or stored, so it's safe for internal prompts and sensitive text.
Important
For planning and estimates only. Prices come from a published rate table dated on the page; providers change pricing without notice, and token counts here are approximations. Confirm against the provider’s own pricing before you budget or commit.
More tools
More from the Hivly network
Free sister tools on our other sites.
