Paper 2606.12191
Agentic Environment Engineering for Large Language Models: A Survey of Environment Modeling, Synthesis, Evaluation, and Application
- Published
- Jun 2026
- Research lab
- Independent
- Citations
- 1
- GitHub
- Not linked
01 In brief
Summary
This paper surveys the field of agentic environment engineering for large language models (LLMs), covering the full lifecycle of environment modeling, synthesis, evaluation, and application.
It defines agentic environments as dynamic, interactive systems for training and evaluating LLM agents, contrasting them with traditional RL simulators.
The survey categorizes environments by eight attributes (e.g., symbolic vs.
neural, open-loop vs.
closed-loop) and eight domains (e.g., GUI, Deep Research, Embodied, Game, Tool, Code, Domain-Specific, Cross-Domain).
For environment synthesis, it distinguishes symbolic synthesis (using code or rules) and neural synthesis (using world models at pixel, word, or latent levels).
Quality evaluation is discussed across correctness, diversity, complexity, and fidelity.
The paper also examines agent evolution through memory-centric, orchestration-centric, trajectory-centric offline, and exploration-centric online methods, and environment evolution via neural-driven, difficulty-driven, and scaling-driven approaches.
Future directions include Environment-as-a-Service, multi-agent environments, neural-symbolic environments, and establishing a science of environment engineering.
02 From the paper
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.