Anthropic has released Claude Opus 5, featuring a new 'effort' parameter that allows users to adjust the trade-off between computational cost and output quality. This parameter, configurable through the Messages API, controls the internal reasoning the model performs before generating a response, with higher effort levels leading to more thinking, higher costs, and longer response times. The 'effort' parameter has five levels, with 'high' being the default, and Anthropic has recalibrated these levels for Opus 5, meaning previous Opus 4.8 settings may not translate directly. Users are advised to test different effort levels to find the optimal balance for their specific use cases, particularly for programming and agent tasks where 'xhigh' is recommended as a starting point. AI
IMPACT Introduces a tunable parameter for cost-performance optimization, potentially impacting enterprise adoption and cost management strategies.
RANK_REASON Frontier-lab model release with new configurable parameter. [lever_c_demoted from frontier_release: ic=1 ai=1.0]
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
