# Cursor Max Mode A mode in [[Cursor.com]] that extends the context window to the maximum the selected model supports, enabling deeper codebase understanding for complex tasks. ## Context Window - Normal mode: ~200k tokens - Max Mode: model's full maximum (varies per model, can be significantly larger) ## How to Enable Open the model selector in chat/agent panel and toggle Max Mode on. Persists as a global setting across conversations. CLI: `/max-mode [on|off]`. Some models are "Max Mode-only" and enable it automatically. ## Pricing - Individual plans (Pro/Pro+/Ultra): 20% upcharge on standard API token rates - Teams/Enterprise: Cursor Token Fee applies instead ## Notes - [[Cursor Cloud Agents]] always run in Max Mode (cannot be disabled) - Reserve for tasks that genuinely need large context (large files, complex cross-file refactors) - For most coding tasks, normal mode works fine ## References - Documentation: https://cursor.com/help/ai-features/max-mode - Changelog: https://cursor.com/changelog ## Related - [[Cursor.com]] - [[Cursor Cloud Agents]] - [[Cursor CLI]]