What it is
The context window, also known as context length, refers to the limit on how much information a large language model (LLM) can "remember" or process in a single interaction. It is measured in tokens, which are typically words or sub-word units. Any information outside this window is effectively forgotten by the model during the current conversation or task, limiting its ability to maintain coherence or draw on distant details from the input. A larger context window allows for more extensive conversations or longer documents.
The size of a model's context window significantly impacts its utility in applications requiring extensive memory, such as summarizing long documents, coding, or maintaining complex conversations. Models with larger context windows often demand more compute resources for both inference and training, influencing AI capex and cloud computing costs. Advances in context window size are a key competitive differentiator among frontier models, enabling more sophisticated and reliable AI interactions for users.
Why it matters
The context window affects how much information an AI can process, impacting its ability to handle long texts or complex conversations, which is crucial for practical use.
Reviewed under editorial standardsUpdated September 26, 2026Not investment advice