Glossary
Context window
How much text a model can consider at once, measured in tokens.
The context window bounds how much document, conversation and instruction the model sees in a single call. Larger windows allow whole documents to be passed directly.
Attention is not uniform across a long context. Material in the middle of a very long input is attended to less reliably than material at either end.
Cost scales with what you put in it, and so does latency. Filling a large window with an entire document set on every call is the most common way teams discover that their AI feature costs more than the problem it solves. Retrieve first, then pass what is relevant.
Commonly misunderstood: A large context window is often treated as a replacement for retrieval. It is usually slower, more expensive and less accurate than retrieving the right few passages.
Related terms, in context
The concepts you almost always meet alongside context window.
- Token
- The unit models read and bill in, roughly three-quarters of an English word, and more for Indian scripts.
- RAG
- Retrieving relevant passages from your own documents and giving them to the model, so answers are grounded and citable.
- Prompt engineering
- Designing the instructions, examples and structure that shape model behaviour.
Where this shows up in our work
Context window is not an abstraction for us. It is a decision we make on live projects. It shows up most directly in llm application development, rag & knowledge retrieval, where getting it wrong has a cost someone can measure.
If you are evaluating a vendor on this, the useful question is not whether they can define the term. It is what they measure, what they would refuse to do, and what happens in their system when the assumption behind context window stops holding.
Questions
What is Context window?
How much text a model can consider at once, measured in tokens.
What do people get wrong about context window?
A large context window is often treated as a replacement for retrieval. It is usually slower, more expensive and less accurate than retrieving the right few passages.
Does Orqent Labs build this?
Yes, LLM Application Development and RAG & Knowledge Retrieval Systems. We work across India, covering all 19,238 PIN codes remotely.
Building something that involves context window?
We will tell you honestly whether it is the right approach for your problem.
Or email bd@dtrasglobal.com · call +91 74118 77878