Context Window
Jendela Konteks
Everything the model sees on one request. Finite, and it decides what fits.
The context window is the total number of tokens the model can attend to in a single request: system prompt, conversation history, files you attached, tool definitions, and the answer it is about to write all share the same budget. Exceed it and the request fails, or the oldest part gets dropped.
Because the model attends to everything in the window equally, adding more text does not simply add more information. It adds competition. The practical consequence: a long document pasted wholesale can push the actual instruction out of the model's focus, which is one of the most common causes of "it ignored my instructions".
Related terms
Token
The unit a model reads and writes. Roughly word-sized, but not exactly.
Context Engineering
Deciding what goes into the model's context, and what stays out.
Attention
The mechanism that lets each token weigh the relevance of every other token.
Prompt
The input you give a model. The word covers more than the question you type.
Memory
A system's ability to carry information from one request to the next.
Want this working inside your team?
It starts with one conversation about your workflow.