Why is my token count higher than the text alone?

Chat APIs wrap every message in structural tokens (role markers, message delimiters) and add a few priming tokens for the reply. Function definitions, tool results, and images all add tokens on top of your visible text, so the billed total is always higher than counting your prompt text alone.

How many tokens does an image cost?

It depends on resolution and detail setting. OpenAI bills images by tiling — a base cost plus per-512px-tile cost at high detail, or a fixed low cost at low detail. Anthropic approximates image tokens as roughly (width x height) / 750. Both treat images as input tokens.

Do function/tool definitions count as tokens?

Yes. The JSON schema for every tool you pass is serialized and counted as input tokens on each request, even if the model does not call the tool. Trimming unused tools and shortening descriptions directly cuts cost.

Are these rules exact?

They are documented approximations that change with model versions. For billing-grade accuracy use the provider's own tokenizer (tiktoken for OpenAI) or the count-tokens endpoint where available.

Does this sheet send anything to a server?

No. It is a static reference rendered in your browser. Selecting a model just filters which rules are shown.

Token Counting Formula Reference Sheet

Token counting, demystified

The token count you are billed for is almost never just your prompt text. Chat APIs add structural overhead, function schemas, tool results, and image tokens on top. This reference sheet collects the counting rules per provider so you can estimate the real billed total before you send a request.

How chat-format token counting works

A chat request is a list of messages, each wrapped in structural tokens:

total = sum(per_message_overhead + content_tokens)
      + tool_schema_tokens
      + image_tokens
      + reply_priming

The per-message overhead (role marker plus delimiters) is a handful of tokens each; the reply priming is a small fixed amount added once. The big, easily forgotten contributors are tool/function schemas (counted every request) and images (counted by resolution), both billed as input tokens.

Tips for accurate estimates

Count tools every time. Every tool definition is re-sent and re-billed on each request — even ones the model never calls. Remove unused tools.
Right-size image detail. Low-detail images cost a small fixed amount; high-detail images scale with resolution. Downscale before sending.
Use the real tokenizer for billing. Approximations are fine for planning; use tiktoken or the provider count-tokens endpoint when accuracy matters.