Skip to main content
Every time you send a message to WonkaChat or an agent, an AI model reads your request and writes an answer. That work is measured in tokens. How tokens are billed depends on your plan.

​
What are tokens?

Tokens are the units that measure AI model usage. Both your question and the AI’s answer use tokens.
  • 1 token ≈ 3/4 of a word in English
  • “Hello, how are you?” ≈ 6 tokens
  • A typical email (200 words) ≈ 265 tokens
  • A full page of text (500 words) ≈ 665 tokens
Files, long conversations, and tool results (for example a web search or a connected app) also add tokens, because the model has to read them.

​
How usage is billed

AI models included

Your seat price includes EU-hosted AI models and €7 in AI credits per seat. There are no separate usage bills.

Your own API key

Your organization connects its own model provider API key. Tokens are billed directly by your provider, at the provider’s rates.
During the 7-day free trial, €5 of AI usage is included.
See Pricing & Plans for seat prices.

​
Why some models use more credits than others

AI models come in different sizes. Larger models are better at complex reasoning, but each token costs more. The models available to you are listed in the model selector in WonkaChat. Your administrator decides which models your organization can use.

​
Tips to use credits well

  • Start with a standard model. Move up to an advanced model only when the answer quality is not good enough.
  • Start a new conversation for a new topic. Long conversations make every new message more expensive, because the model re-reads the history.
  • Attach only the files you need. Every attached document is read by the model.
  • Give each agent the right model. A simple FAQ agent does not need the most advanced model.

​
Frequently asked questions

Yes. You can pick a model for each conversation, and each agent can use a different model. There is no penalty for switching.
The current model is shown in the model selector at the top of the conversation. For an agent, the model is set in the agent’s settings.
Your WonkaChat administrator sees your organization’s usage. For billing questions, contact support.