Glossary

Context Length

Context length is the maximum number of tokens—representing characters, words, or sub-word units—that a large language model can process in a single interaction. It defines the total capacity of the model's input and output window, encompassing both the user's prompt and the model's generated response within a single session.

Context length determines the depth and breadth of information a model can analyze at once. As AI agents increasingly manage complex workflows, such as analyzing entire codebases or multi-channel marketing histories, a larger context window allows for greater coherence and accuracy. When input data exceeds this limit, the model must truncate information or rely on external retrieval systems, which can lead to the loss of critical instructions or historical data necessary for maintaining long-term task continuity.

In practice, managing context length requires balancing input volume against performance and cost. Practitioners should monitor token usage to ensure essential instructions remain within the active window, as models may prioritize newer information over earlier inputs. When working with extensive datasets, developers often implement techniques like sliding windows or Retrieval-Augmented Generation (RAG) to inject relevant information dynamically. Monitoring these constraints is essential for optimizing agent reliability and ensuring that automated processes remain aligned with established project parameters.

Last updated: 2026-09-02