CLAUDE AGENTS GET 'DREAMING' ABILITY

The new feature enables Claude Managed Agents to process information more deeply before responding, improving their ability to handle complex tasks. The internal reasoning step works by having agents generate extended thoughts before taking actions, allowing for more nuanced problem-solving. Anthropically describes the capability as agents being able to 'dream'—a simplified way of explaining how the system can pause and reflect on problems before proceeding. Separately, Anthropic increased the usage limits for Claude Code from 5 hours to 10 hours monthly for both Pro and Max subscribers. The enhancement aims to give developers and power users more flexibility when working with Claude's coding capabilities. These updates represent incremental improvements to Claude's agent functionality and usage tiers, keeping pace with competition in the AI assistant market.