SESKStuff Everyone Should Know
SESK: Stuff Everyone Should Know · sesk.info/ai-api-billed-by-tokens

LLM APIs usually charge per request based on what?

The number of input and output tokens.

Start playing free

  1. The number of words the user types
  2. Minutes connected to the server
  3. Input tokens plus output tokens ✔

Why: You pay for the tokens you send in, including instructions and tool definitions, and the tokens the model writes back. Long prompts and long answers both cost more, and both add latency too.

Huh, didn't know that: On the Claude API, adding tools also adds a special tool use system prompt of a few hundred tokens to each request.

The dad joke
AI charges by the token. My kids charge by the chore, and the rates keep going up.
Like this card?
Play the game

References

Answer: Input tokens plus output tokens. You pay for the tokens you send in, including instructions and tool definitions, and the tokens the model writes back. Long prompts and long answers both cost more, and both add latency too.

  1. Claude API Docs, Tool use with Claude platform.claude.com/docs/en/agents-and-tools/tool-use/overview
    The total number of input tokens sent to the model (including in the `tools` parameter) The number of output tokens generated
    Checked 2026-10-11.
  2. Claude API Docs, Tool use with Claude platform.claude.com/docs/en/agents-and-tools/tool-use/overview
    the API also automatically includes a special system prompt for the model that enables tool use
    Checked 2026-10-11.

Think this answer is wrong? Tap "Challenge this answer" on the card in the game and tell us why.

Test yourself in the game

Artificial IntelligenceWhat does a learned embedding assign to each input token?Artificial IntelligenceWhich statement correctly distinguishes model parameters from tokens?Artificial IntelligenceWhy is one written word not always one language-model token?

That's stuff everyone should know.