Glossary · AI
What is
Context Window?
The maximum number of tokens an LLM can process in a single call.
By Anish· Founder · Vedwix
·Definition
The context window is the model's working memory — the total tokens (prompt + response) it can handle in one inference. Modern frontier models support 200k to 2M tokens, but quality often degrades past a few hundred thousand. Long-context use cases include document analysis, codebase reasoning, and multi-turn agents.
Example
Claude's 200k context window can hold roughly 500 pages of text, enough for a full legal contract or medical record.
How Vedwix uses Context Window in client work
We design RAG over long-context whenever the corpus exceeds 100k tokens — chunking and retrieval still beat brute-force context.
Building with Context Window?
We ship this.
If you're building with Context Window in production, we can help — from architecture review to full implementation.
Brief us