Knowledge base
AI models
A token is a unit used to measure the resources consumed when an AI model processes information and generates responses.
- A single token may represent a complete word, part of a code expression, or another sequence of characters.
- Token consumption depends on the text, language, and AI model being used.
- Armenian and English texts of similar length may contain different numbers of tokens.
- Most AI services calculate usage costs based on the number of input tokens processed and output tokens generated.
17.10.2026
The cost of using an AI model is calculated based on the volume of information sent to the model (input tokens) and the responses it generates (output tokens).
- Token pricing is typically quoted per one million tokens.
- Rates can vary significantly depending on the AI model.
- Output tokens generally cost more than input tokens.
- Different pricing methods may apply to audio, image, and video processing.
17.10.2026
Last update 19.10.2026