Agentic Large Language Models, a survey

2025-03-29

AI Summary

This paper addresses the fragmentation of capabilities and heavy data dependency hindering the development of embodied agentic large language models (LLMs). We propose a unified “Reasoning–Acting–Interacting” tri-dimensional framework to systematically characterize their synergistic mechanisms. Methodologically, we integrate chain-of-thought reasoning, reflection mechanisms, tool invocation, multi-agent simulation, retrieval-augmented generation (RAG), and role-aware interaction modeling. We first reveal the performance gains from retrieval-augmented tool use and reflection-driven multi-agent collaboration, and introduce the “reasoning-as-learning” paradigm to mitigate training data scarcity. Our contributions include: (1) establishing the first tri-dimensional taxonomy that clarifies technical trajectories and core challenges; (2) identifying high-impact application domains—medical diagnosis, financial analysis, and scientific discovery; and (3) empirically demonstrating that self-reflective agents can restructure scientific workflows and enable continual learning.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Multiagent Systems: Modeling other Agents
• Cognitive Modeling & Cognitive Systems: Agent Architectures

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Search and Retrieval-Augmented AI: Agentic search
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation

Abstract

There is great interest in agentic LLMs, large language models that act as agents. We review the growing body of work in this area and provide a research agenda. Agentic LLMs are LLMs that (1) reason, (2) act, and (3) interact. We organize the literature according to these three categories. The research in the first category focuses on reasoning, reflection, and retrieval, aiming to improve decision making; the second category focuses on action models, robots, and tools, aiming for agents that act as useful assistants; the third category focuses on multi-agent systems, aiming for collaborative task solving and simulating interaction to study emergent social behavior. We find that works mutually benefit from results in other categories: retrieval enables tool use, reflection improves multi-agent collaboration, and reasoning benefits all categories. We discuss applications of agentic LLMs and provide an agenda for further research. Important applications are in medical diagnosis, logistics and financial market analysis. Meanwhile, self-reflective agents playing roles and interacting with one another augment the process of scientific research itself. Further, agentic LLMs may provide a solution for the problem of LLMs running out of training data: inference-time behavior generates new training states, such that LLMs can keep learning without needing ever larger datasets. We note that there is risk associated with LLM assistants taking action in the real world, while agentic LLMs are also likely to benefit society.

Problem

• Surveying agentic LLMs' reasoning, action, and interaction capabilities
• Exploring applications in medicine, logistics, and financial analysis
• Addressing data scarcity via inference-time learning in LLMs

Innovation

• Agentic LLMs reason, act, and interact
• Multi-agent systems enable collaborative tasks
• Self-reflective agents enhance scientific research