Learn
What is a context window?
Last reviewed July 16, 2026
A context window is the maximum amount of text a model can consider in one request, measured in tokens (roughly three-quarters of a word each). Everything counts against it: your message, the conversation history, attached documents, and the model's own reply. Past the limit, something must be dropped, summarized, or refused.
What fits in practice
A 128K-token window holds roughly a 300-page book; a 1M-token window holds a small codebase or a stack of papers. But models attend unevenly across very long contexts, so more window is not automatically more comprehension: retrieval quality in the middle of a huge context is a known weak spot across model families.
Context and cost
Input tokens are billed, so a long context is a recurring cost: every message in a long conversation re-sends the history. Practical systems manage this with compaction (summarizing older turns), selective attachment (only the relevant files), and caching where providers offer it.
Every idapt model page lists its context window and per-token prices; long chats compact automatically, and the pre-send estimate shows what a message will cost before it goes.
Frequently asked
See it in practice
The fastest way to understand it is to use it. Start free, no credit card.
Try idapt free