Agentic Environment Engineering for Large Language Models: A Survey of Environment Modeling, Synthesis, Evaluation, and Application

2026-06-10

AI Summary

This study addresses the lack of systematic classification and analysis of environments in current large language model (LLM) agent research, which hinders the co-evolution of agents and their environments. The work proposes the first environment engineering framework tailored for LLM agents, introducing a taxonomy based on eight core attributes and domains. It integrates symbolic and neural synthesis methods to enable automated environment generation and employs a multidimensional evaluation mechanism to support dynamic, scalable agent–environment interaction studies. Furthermore, the paper identifies four archetypal agent evolution pathways and three environmental evolution paradigms, mapping the capability spectrum of existing environment systems. This provides both theoretical foundations and practical guidance for agent training, evaluation, and deployment, advancing environment engineering toward standardization and service-oriented architectures such as Environment-as-a-Service.

Abstract

Environments serve as interactive systems for large language model (LLM) based agents across diverse scenarios and play a crucial role in driving the continual evolution of model capabilities. Despite this importance, existing work lacks a systematic categorization and deep analysis. This paper systematically studies current researches on agentic environments from the perspective of the environment engineering lifecycle, covering their modeling, synthesis, evaluation and application. Specifically, the paper first introduces representative environments from the perspectives of eight attributes and eight domains, providing detailed analyses of their development paths and highlighting their core capabilities. Second, for automated environment synthesis, two paradigms are introduced, such as symbolic synthesis and neural synthesis. This paper also shows different environment evaluation methods in each paradigm. Thirdly, the corresponding environment applications from the perspective of agent-environment co-evolution are discussed. In specific, the paper characterizes the primary pathways for agent evolution in dynamic environments from four complementary perspectives: memory-centric experience evolution, orchestration-centric workflow evolution, trajectory-centric offline evolution, and exploration-centric online evolution. And three paradigms of environment evolution are identified, namely neural-driven, difficulty-driven, and scaling-driven approaches. At last, several promising future directions are discussed, including Environment-as-a-Service, Multi-agent Environments, and Neural-Symbolic Environments.

Problem

• Agentic Environment
• Large Language Models
• Environment Engineering
• Agent-Environment Interaction
• Environment Evaluation

Innovation

• Agentic Environment Engineering
• Environment Synthesis
• Agent-Environment Co-evolution
• Neural-Symbolic Environments
• Environment Evaluation