Cook with AI Enrollment closed
AI & LLMs Runs locally

Token Counter & Prompt Cost

Paste a prompt, get an approximate token count and what it costs as input across every major LLM.

ELI5

It tells you roughly how big one prompt is and what sending it costs.

A little more detail

What this tool does

Providers price model input by token. Paste a system prompt, user message, or retrieved context to estimate its token count and input cost across several models.

Use it to
  • Check the cost of a long system prompt
  • Estimate input cost for one call or a large batch
  • Compare the same prompt across model providers
Interactive workspace Results update as you type
Your prompt

Token count uses a chars÷4 heuristic (~±15% vs real tokenizers). Actual counts vary by model and language.

Or enter token count directly
Disabled while prompt text is filled — clear the text to use this field.
Estimate

Paste text or enter a token count to see costs.

Cost by model

Input tokens only, standard API tier. Sorted cheapest first.

Model Provider Tier Cost (this prompt) Cost per 1k runs Cost per 1M runs

Your text never leaves the browser. Token counts and costs are computed entirely on your device — nothing is sent to a server.

When you paste text, the tool estimates tokens with a simple chars÷4 rule — roughly one token per four characters of English prose. Real tokenizers split differently: code, JSON, and non-English text often use more tokens per character.

Costs come from each provider's published input price per million tokens. The table shows what you'd pay to send this prompt once, then scaled to 1,000 and 1,000,000 identical runs. Output tokens, prompt caching, and batch discounts are not included — use the inference cost calculator for full monthly estimates.