This research surveys the emerging field of agentic large language models (LLMs) - AI systems that can reason, act, and interact with their environment.

• Reasoning capabilities: LLMs that can reflect on their decision-making and retrieve external knowledge
• Action mechanisms: Integration with tools, robots, and action models for real-world impact
• Interactive systems: Agents that can engage with users and other systems adaptively

In medical applications, agentic LLMs show promise for assisting with diagnosis while raising important security considerations as these systems gain the ability to take autonomous actions.

Agentic Large Language Models, a survey