Anthropic's latest Claude models represent a step change in the horizontal capabilities that matter most for healthcare: deeper reasoning over complex clinical data, persistent memory across long patient interactions, and agentic search that connects to authoritative medical sources. But general capability is only the starting point. In this talk, we'll walk through recent advancements in Claude's core reasoning, memory, and research capabilities, before showing how "agent skills" enable builders to transform Claude from a powerful generalist into a genuine healthcare expert.