Context window

A context window is the limited amount of tokenized information a model can consider together during one generation or inference.

Plan for this page

Understand one concept in plain language and connect it to nearby ideas.

You will leave with: It can include instructions, conversation, documents, tool results, and generated tokens, depending on the product. Information outside the window is not automatically available unless another system retrieves, summarizes, or stores it.
Time: 3–8 min
Before you begin: No prerequisite. Keep one real example in mind.
Do this now: Read the explanation, then compare the example with the common misunderstanding.

Prepared next stepCompare context, retrieval, and tools

Context windowA context window is the limited amount of tokenized information a model can consider together during one generation or inference.

Token
Prompt
System instruction
Memory

Plain explanation

It can include instructions, conversation, documents, tool results, and generated tokens, depending on the product. Information outside the window is not automatically available unless another system retrieves, summarizes, or stores it.

Example

A long report may need to be selected or divided because the report, question, instructions, and expected answer must fit the available context.

Why it matters

A larger window can admit more information but does not guarantee that every detail is used correctly. Relevance, position, conflicts, retrieval quality, privacy, and cost still matter.

Common misunderstanding

Misconception: the model remembers every previous conversation forever. Correction: immediate context is bounded, while any durable memory is a separate system and policy.

Related concepts

Follow relationships, not only the alphabet

TokenA token is a unit into which a model’s input or output is divided for processing.PromptA prompt is input supplied to a generative model or system to shape the task, context, constraints, or desired output.System instructionA system instruction is configuration context intended to guide a model’s behavior across a session or task, often with priority over ordinary user input.MemoryMemory is information a system retains or can retrieve for use beyond the immediate step or context.

Compare context, retrieval, and tools

Sources and scope

What supports this object

Machine Learning GlossaryGoogle for Developers

Defines context windows for language models; provider limits and allocation between input and output are model- and version-specific.

Open source

Loading page content...

…