Zing Forum

Forum HomeLLM Answers & Content StrategyReading

Reading

Panoramic Research on LLM Game Agents: An Academic Resource Repository from Minecraft to Multi-Agent Collaboration

The awesome-LLM-game-agent-papers repository maintained by the DISL Lab at Georgia Tech systematically organizes research on LLM-based game agents, covering 8 major domains such as Minecraft, text adventures, and social reasoning. It includes over 500 papers and has been formally accepted for publication in ACM Computing Surveys.

LLM游戏智能体Minecraft多智能体强化学习具身智能文本冒险社交推理ACM CSUR
Published 2026-05-22 09:10Recent activity 2026-05-22 09:19Estimated read 5 min
Panoramic Research on LLM Game Agents: An Academic Resource Repository from Minecraft to Multi-Agent Collaboration
1

Section 01

Introduction: Overview of the Panoramic Resource Repository for LLM Game Agent Research

The awesome-LLM-game-agent-papers repository maintained by the DISL Lab at Georgia Tech systematically organizes research on LLM-based game agents, covering 8 major domains such as Minecraft, text adventures, and social reasoning. It includes over 500 papers, and its corresponding survey paper of the same name has been formally accepted by ACM Computing Surveys. This resource repository provides a panoramic navigation from entry-level to in-depth content for researchers and developers.

2

Section 02

Background: Games as Ideal Testbeds for AI Research

Games have always been ideal testbeds for AI research—from Deep Blue defeating the chess champion to AlphaGo conquering Go, and now to LLM agents demonstrating reasoning and collaboration capabilities in open-world games, the field is advancing rapidly. The awesome-LLM-game-agent-papers repository is an authoritative resource compilation in this domain, with systematic classification and comprehensive content.

3

Section 03

Core Research Domains: 8 Major Directions and Key Achievements

The resource repository divides into 8 core domains:

  1. Minecraft (Embodied intelligence research such as Voyager, JARVIS-1)
  2. Text adventures (Pure language reasoning like ReAct, Reflexion)
  3. Social reasoning (Multi-agent interaction such as WarAgent, Richelieu)
  4. Competitive games (Strategic research on chess, StarCraft)
  5. Multi-agent collaboration (Coordination frameworks like VillagerAgent, CausalMACE)
  6. Simulated societies (Social agent architectures and emergent behaviors)
  7. Embodied simulation (VLA models and physical world representation)
  8. Technical foundations (Supporting technologies like planning, memory, training)
4

Section 04

Technical Evolution: Three Generations of Development

The technical evolution of the domain is divided into three generations:

  1. 2022-2023: Establishment of basic architectures (ReAct paradigm, Voyager's code generation capabilities)
  2. 2023-2024: Multimodal and memory enhancement (GPT-4V-related visual fusion, JARVIS-1 memory framework)
  3. 2024-2025: Collaboration and self-evolution (TeamCraft collaboration framework, RetroAgent's autonomous learning)
5

Section 05

Practical Value: Cross-domain Application Prospects

The research results have multi-domain application value:

  • Game industry: Intelligent NPCs, automated testing, content generation
  • Robotics and embodied intelligence: Virtual training grounds, transfer of planning and decision-making
  • Multi-agent systems: Foundations for autonomous driving, improved robustness of negotiation systems
6

Section 06

Research Challenges: Key Issues to Be Addressed

The current domain faces four major challenges:

  1. Scalability of long-term planning (Loss of direction in ultra-long tasks)
  2. Physical common sense grounding (Decisions violating common sense)
  3. Multi-agent alignment (Collaboration interference and safety)
  4. Standardization of evaluation (Lack of unified benchmarks)
7

Section 07

Conclusion and Recommendations

awesome-LLM-game-agent-papers is a panoramic map of the domain, outlining the potential of LLM general agents. It is recommended that researchers start with classic papers like ReAct and Voyager; developers can draw on the technologies to build the next generation of game AI. The repository is open to paper contributions (submit issues or PRs).

Continue Reading

Keep going with more reads from the same topic.

01
Nornir MCP Server: An Enterprise-Grade Bridge for Integrating Large Language Models into Network Automation

Nornir MCP Server: An Enterprise-Grade Bridge for Integrating Large Language Models into Network Automation

Nornir MCP Server is an enterprise-level server based on the Model Context Protocol (MCP). It seamlessly integrates large language models (such as Claude) with the Nornir network automation framework, supporting natural language orchestration for multi-vendor network devices (Cisco, Arista, Juniper, etc.), and providing production-grade features like a dual-engine architecture (NAPALM + Netmiko), intelligent filtering, and a secure sandbox.

Recent activity 2026-05-06 20:51
02
Bibliothèque Française LLM: A French Public Domain Literature Index System Optimized for Large Language Models

Bibliothèque Française LLM: A French Public Domain Literature Index System Optimized for Large Language Models

Bibliothèque Française LLM is a structured indexing and annotation project for French public domain literature designed specifically for large language models (LLMs). It integrates multiple authoritative sources such as DraCor, Common Corpus, and Wikisource, providing metadata indexing categorized by genre, author, and era, as well as in-depth annotations for dramatic texts (including characters, lines, stage directions, etc.). Its aim is to enable LLMs to efficiently read and understand classic French literary works.

Recent activity 2026-05-06 20:50
03
Splinter: A Lock-Free Zero-Copy Shared Memory KV and Vector Storage Library That Eliminates Socket and Memcpy Overhead for LLM Inference

Splinter: A Lock-Free Zero-Copy Shared Memory KV and Vector Storage Library That Eliminates Socket and Memcpy Overhead for LLM Inference

Splinter is a minimalist, high-performance key-value (KV) and vector storage system enabling zero-latency inter-process communication via shared memory and atomic operations. With only 766 lines of core code, it supports millions of operations per second and 768-dimensional vector storage, offering a new architectural approach for local LLM inference and data-intensive applications.

Recent activity 2026-04-03 08:49
04
Folkering OS: When the Operating System Itself Is AI—A Self-Evolving Bare-Metal Rust System

Folkering OS: When the Operating System Itself Is AI—A Self-Evolving Bare-Metal Rust System

Folkering OS is the world's first AI-native bare-metal operating system, entirely written in Rust no_std without relying on Linux, POSIX, or libc. It can generate commands from scratch, compile them into WASM, and run them in 10 seconds, achieving true self-evolution.

Recent activity 2026-04-09 16:15