mlsanityData to 7 Oct 2026

Lumine: An Open Recipe for Building Generalist Agents in 3D Open Worlds

arXiv 2511.08892 ↗First posted 12 Nov 2025cs.AIcited by 19

On the map: LLM agents and tool use → Reinforcement learning for LLM agents

Abstract (arXiv)

We introduce Lumine, the first open recipe for developing generalist agents capable of completing hours-long complex missions in real time within challenging 3D open-world environments. Lumine adopts a human-like interaction paradigm that unifies perception, reasoning, and action in an end-to-end manner, powered by a vision-language model. It processes raw pixels at 5 Hz to produce precise 30 Hz keyboard-mouse actions and adaptively invokes reasoning only when necessary. Trained in Genshin Impact, Lumine successfully completes the entire five-hour Mondstadt main storyline on par with human-level efficiency and follows natural language instructions to perform a broad spectrum of tasks in both 3D open-world exploration and 2D GUI manipulation across collection, combat, puzzle-solving, and NPC interaction. In addition to its in-domain performance, Lumine demonstrates strong zero-shot cross-game generalization. Without any fine-tuning, it accomplishes 100-minute missions in Wuthering Waves and the full five-hour first chapter of Honkai: Star Rail. These promising results highlight Lumine's effectiveness across distinct worlds and interaction dynamics, marking a concrete step toward generalist agents in open-ended environments.

Key facts. Cited by 19 papers in the map (data through 7 Oct 2026). Placed in Reinforcement learning for LLM agents, one of 16 directions in LLM agents and tool use. First posted on arXiv 12 Nov 2025.

Builds on

0 of 0 references.

No references here yet.

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2025: 2; 2026: 17.

19 papers.

Sort:

By area:

  1. 2025SIMA 2: A Generalist Embodied Agent for Virtual Worldscited by 26
  2. 2026NitroGen: An Open Foundation Model for Generalist Gaming Agentscited by 22
  3. 2026GameWorld: Towards Standardized and Verifiable Evaluation of Multimodal Game Agentscited by 22
  4. 2026Scaling Behavior Cloning Improves Causal Reasoning: An Open Model for Real-Time Video Game Playingcited by 10
  5. 2026GameVerse: Can Vision-Language Models Learn from Video-based Reflection?cited by 7
  6. 2026Research on World Models Is Not Merely Injecting World Knowledge into Specific Taskscited by 3
  7. 2026MapTab: A Diagnostic Benchmark for Long-Horizon Multi-Criteria Multimodal Reasoning on Heterogeneous Topological Graphscited by 3
  8. 2026Hera: Learning Long-Horizon Coordination for Device-Cloud Collaborative LLM Agentscited by 3
  9. 2025IPR-1: Interactive Physical Reasonercited by 2
  10. 2026MAIN-VLA: Modeling Abstraction of Intention and eNvironment for Vision-Language-Action Modelscited by 2
  11. 2026Odysseus: Scaling VLMs to 100+ Turn Decision-Making in Games via Reinforcement Learningcited by 2
  12. 2026Towards Generalist Game Players: An Investigation of Foundation Models in the Game Multiversecited by 2
7 more
  1. 2026SPIKE: An Adaptive Dual Controller Framework for Cost-Efficient Long-Horizon Game Agentscited by 2
  2. 2026SimpleTool: Parallel Decoding for Real-Time LLM Function Callingcited by 1
  3. 2026ActSWM: Action-Sensitive World Models for Long-Horizon Planning in Open-World Gamescited by 1
  4. 2026UrbanGround: From Local Perception to Spatial Agency in a Real-Scale Citycited by 1
  5. 2026GROW: Aligning GRPO with State-Action Modeling for Open-World VLM Agentscited by 0
  6. 2026GameWAM: A World Action Model for Video Gamescited by 0
  7. 2026GameHorizon Suite: Multi-Horizon Data and Evaluation in Gameplaycited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026MineExplorer: Evaluating Open-World Exploration of MLLM Agents in Minecraft
  2. 2026OpenGameEval: Benchmarking Agentic Programming and Exploration in a Stateful Game Engine
  3. 2023JARVIS-1: Open-World Multi-task Agents with Memory-Augmented Multimodal Language Models
  4. 2024Craftium: Bridging Flexibility and Efficiency for Rich 3D Single- and Multi-Agent Environments
  5. 2025Game-TARS: Pretrained Foundation Models for Scalable Generalist Multimodal Game Agents
  6. 2026GS-Agent: Creating 4D Physical Worlds With Generative Simulation
  7. 2025OmniActor: A Generalist GUI and Embodied Agent for 2D&3D Worlds
  8. 2025OpenHA: A Series of Open-Source Hierarchical Agentic Models in Minecraft