Claude Code has recently received a series of updates that significantly enhance developer efficiency when coding with Claude AI. These changes center on increasing the output token limit, introducing a loop task command, and reducing costs — improvements that directly affect the smoothness of your workflow if you rely on Claude for daily programming.
Default Output Token Increased to 64K
In the latest version, the default maximum output token count for the Opus 4.6 model has been raised to 64K. This means developers no longer need to manually set higher output limits to handle longer code generation or refactoring tasks. Complex outputs that previously required batching can now be obtained in a single response, reducing the waiting time from multiple calls. This improvement is the result of ongoing optimizations to the underlying Claude AI model, making long-form text generation more practical.
New /loop Command for Repetitive Tasks
The newly introduced /loop command is a standout feature in this update. It allows developers to repeatedly execute a specific action within the same context — for example, batch-modifying code structures or generating multiple functions in sequence. This feature is ideal for repetitive tasks that require maintaining context consistency. Combined with Claude AI's strong understanding capabilities, the loop command effectively reduces manual adjustments and increases automation.
