Planning in the LLM Era: Building for Reliability and Efficiency
For researchers and practitioners building intelligent agents, this paper provides a perspective on how to leverage LLMs for planning in a way that balances reliability and efficiency.
The paper argues that the planning field is realigning in the LLM era, shifting from single-shot or hybrid LLM-based planning towards generating symbolic planners that are verified and efficient at inference time. It categorizes planner-generation methods, discusses their limitations, and outlines research steps for more reliable and efficient LLM-based planner generation.
Growing attention to intelligent agents has put a spotlight on one of their central capabilities: planning. Early attempts to leverage large language models (LLMs) for planning relied on single-shot plan generation, followed by hybrid approaches that coupled LLMs with limited external search. These methods, unsound and incomplete by their very nature, often require substantial resources without yielding better solutions on unseen problems. As the limitations of LLMs become clearer, recent work has shifted toward using them at solution construction time -- generating symbolic solvers for a family of problems that can be verified and then used efficiently at inference time. This trend reflects the growing need for agents that are both reliable and resource-efficient. It also offers a path towards generating maintainable planners with minimal dependence on language models at inference time. In this paper, we argue that this shift reflects a broader realignment of the planning field in the LLM era. We examine three major categories of planner-generation methods, discuss their current limitations, and outline research steps towards a more reliable and efficient LLM-based generation of planners.