According to Claude on Twitter, the Claude Platform now supports an advisor–executor pattern where Opus serves as a strategic advisor while Sonnet or Haiku executes tasks, enabling near Opus-level reasoning at a fraction of the cost. As reported by the original tweet from @claudeai on April 9, 2026, teams can architect agents that route complex planning and critique to Opus and delegate high-volume generation or tool use to Sonnet or Haiku, creating a cost-efficient, scalable agent workflow. According to Anthropic’s Claude product positioning, Opus provides strongest reasoning, while Sonnet and Haiku deliver faster, cheaper execution, which implies clear business opportunities in support automation, AI operations, and software agents where unit economics are sensitive to model spend.