token
The unit an LLM processes — a sub-word chunk produced by the model's tokenizer, billed individually.
Definition
A token is the unit an LLM ingests and emits — a sub-word fragment produced by the model's tokenizer (BPE for GPT, sentencepiece-ish for Claude). English text averages roughly four characters per token. All LLM pricing and rate limits are denominated in tokens. Counting tokens before sending long prompts is essential for budgeting.
When to use
See also
- context window — The maximum number of tokens an LLM can consider at once — input prompt plus output combined.
- LLM — Large Language Model — a transformer-based model trained on internet-scale text to generate and reason.