Anthropic multiagent sessions and Outcomes enter public beta

AI Impact Summary

Anthropic is releasing multiagent sessions and Outcomes as public beta features, accessible via the managed-agents-2026-04-01 beta API header. Multiagent sessions enable coordinated workflows across multiple Claude instances with shared context and state management, while Outcomes provide structured result validation and routing. Teams building complex agentic systems can now test these capabilities in production-like environments before general availability, though beta APIs may change and should not be used for mission-critical workloads without fallback plans.