Context window
The maximum amount of text a model can read in one request, measured in tokens.
What it means
A context window is the maximum amount of text a model can read in one request, measured in tokens. Search results count against it, along with the prompt and the model's reply. Long search results can crowd out instructions or push earlier content out of the window, which lowers answer quality even when the model supports a large window.
Search APIs help manage this through budgets. Some let you set a token limit for the results, cap the content per page, or choose how many snippets come back. OpenAI's web search tool offers a context size setting with low, medium and high options. Count tokens before you ship, and keep the budget well below the model's limit so there is room for the answer.
Continue this reading path
Cite this entry
https://bestwebsearchapi.com/glossary/context-window/