~/CLAUDE API/anthropic-adds-mid-conversation-effort-adjustment-feature-to-claude-api

Anthropic Adds Mid-Conversation Effort Adjustment Feature to Claude API

Anthropic has introduced a beta feature for the Claude API that allows developers to dynamically adjust the model's effort level mid-conversation. Developers can modify the setting without altering top-level configurations, enabling smooth transitions between casual chat and complex analysis. This feature helps developers optimize response thoroughness, speed, and cost dynamically within a single workflow. By allowing mid-conversation adjustments without breaking prompt caching, it improves both operational efficiency and cost-effectiveness for application builders. Developers can adjust the effort tier by inserting an empty system message with the `output_config.effort` setting into the message sequence. The change takes effect starting from the next user message containing the new input, influencing thinking, tool calls, and prose generation.

## BACKGROUND

Prompt caching is a technique used by LLM providers to store and reuse frequently accessed prompt contexts, significantly reducing latency and API costs. Reasoning effort parameters control how many tokens a model allocates for internal thinking and processing before responding, balancing execution speed and depth.

## REFERENCES

## KEYWORDS

#Claude API#Anthropic#AI Development#LLM Optimization#API Feature

$ subscribe --daily