Definition·
Fundamentals

Context Window

The maximum number of tokens a model can consider at once across input and output.

Detailed explanation

The context window decides what an LLM can "see": conversation, attached documents, system instructions. The larger it is, the more content you can inject (e.g. RAG), at higher cost and latency. Recent models reach hundreds of thousands of tokens.

Examples

Claude 3.5 Sonnet: 200k tokens
Gemini 1.5 Pro: 1M+ tokens
GPT-4 Turbo: 128k tokens

Frequently asked questions

Is bigger always better?

Not always: past a point recall quality drops and cost rises. RAG remains useful.

Related terms

Last updated: 7/15/2026

Talent AI

Turn theory into practice

Post a mission or join the community of top AI, Data and Machine Learning experts.