Tune Model Verbosity for Specific Professional Contexts
Control Sonnet 5's high verbosity with explicit length constraints.
Use specific constraints in your prompts to manage Sonnet 5’s inherent verbosity, ensuring responses match your required level of detail.
The Scenario
You need a quick summary of a long meeting transcript to share on Slack, but the AI keeps providing a 10-paragraph essay that no one will read.
Before & after
Users often accepted default outputs and manually edited them to fit the required length, costing 5–10 minutes per response for editing and formatting.
Explicitly state your desired verbosity in the system prompt. In 30 seconds, you can restrict or expand the response style, saving you from reading extraneous text.
The Prompt
Task: [DESCRIBE_TASK]. Constraints: Provide a [CONCISE_OR_DETAILED] response. Skip non-essential preamble and focus only on the [SPECIFIC_OUTPUT]. Use a [PROFESSIONAL_OR_CASUAL] tone.
Sonnet 5 calibrates length based on task complexity but is noted for being 'very verbose.' If you need short, punchy updates for Slack or long, detailed reports, you must be explicit in your prompt instructions.
Source
Claude Platform - Claude Platform Docs"Claude Sonnet 5 calibrates response length to the complexity of the task... If your product depends on a certain style... you may need to tune your prompts."