Every model has a hard ceiling on how many tokens it can consider at once. Here is what counts toward it, every current limit, and how to stay under it.
LLM Token Counter — open the live counter