A: Hi there

Tokens in an LLM are calculated based on three factors:
1. Input: The query provided by the user.
2. Context: Additional information like files, searches, or knowledge bases attached to the query.
3. Output: The model's final response.
Simple queries with no added context consume fewer tokens. When context is attached, the system processes this data—searching, retrieving, and preparing relevant information. These steps, handled by the LLM, increase token usage proportionally.

This ensures responses are accurate and tailored to the user's needs while maintaining clarity and efficiency.
Regards