Steer Model Reasoning with the New Effort Parameter
Match AI response depth to task complexity using effort levels.
Use the 'effort' parameter to toggle how much the model 'thinks' before answering, ensuring high accuracy for hard tasks and efficiency for easy ones.
The Scenario
You need to balance the AI's 'thinking' depth with the complexity of your task to ensure you get the most accurate result for hard problems or the fastest for easy ones.
Before & after
Using the same high-power prompt for every task, leading to overly verbose answers for simple questions or under-thought answers for complex ones. Total time: Constant re-prompting.
Set the 'effort' parameter to 'max' for capability-critical judgment or 'low' for simple formatting to save time/cost. Total time: 1 minute to adjust settings.
The Prompt
[TASK_DESCRIPTION]. Set effort to [LOW/MEDIUM/HIGH/MAX] because this task is [ROUTINE/COMPLEX/CAPABILITY-CRITICAL]. Give me the output in [FORMAT].
The new 'effort' parameter is the primary way to steer Opus 5. 'Max' is for difficult engineering, while 'low' reduces verbosity and thinking time for routine tasks.
Source
Claude Platform release notes - Claude Platform Docs"Effort is the primary control for steering Claude Opus 5... with max for capability-critical work."