Level Generation Through Large Language Models
Not on the map. Why?
Abstract (arXiv)
Large Language Models (LLMs) are powerful tools, capable of leveraging their training on natural language to write stories, generate code, and answer questions. But can they generate functional video game levels? Game levels, with their complex functional constraints and spatial relationships in more than one dimension, are very different from the kinds of data an LLM typically sees during training. Datasets of game levels are also hard to come by, potentially taxing the abilities of these data-hungry models. We investigate the use of LLMs to generate levels for the game Sokoban, finding that LLMs are indeed capable of doing so, and that their performance scales dramatically with dataset size. We also perform preliminary experiments on controlling LLM level generators and discuss promising areas for future work.
Key facts. Cited by 68 papers in the map (data through 14 Oct 2026). Builds on 22 of its 35 references that are in the map. First posted on arXiv 11 Feb 2023.
Builds on
22 of 35 references.
References · 22
- 2020Language Models are Few-Shot Learnerscited by 32,419
- 2020Scaling Laws for Neural Language Modelscited by 6,586
- 2022Hierarchical Text-Conditional Image Generation with CLIP Latentscited by 6,585
- 2021Decision Transformer: Reinforcement Learning via Sequence Modelingcited by 1,546
- 2021Calibrate Before Use: Improving Few-Shot Performance of Language Modelscited by 1,392
- 2022A Generalist Agentcited by 808
- 2021CodeXGLUE: A Machine Learning Benchmark Dataset for Code Understanding and Generationcited by 744
- 2023MusicLM: Generating Music From Textcited by 448
- 2017Imagination-Augmented Agents for Deep Reinforcement Learningcited by 140
- 2017Procedural Content Generation via Machine Learning (PCGML)cited by 137
- 2020PCGRL: Procedural Content Generation via Reinforcement Learningcited by 90
- 2020Deep Learning for Procedural Content Generationcited by 73
10 more
- 2016Super Mario as a String: Platformer Level Generation Via LSTMscited by 40
- 2021Learning Controllable Content Generatorscited by 40
- 2019Bootstrapping Conditional GANs for Video Game Level Generationcited by 35
- 2020Griddly: A platform for AI research in gamescited by 31
- 2020Interactive Evolution and Exploration Within Latent Level-Design Space of Generative Adversarial Networkscited by 22
- 2020Conditional Level Generation and Game Blendingcited by 17
- 2017Learning to Decode for Future Successcited by 16
- 2020Multi-Domain Level Generation and Blending with Sketches via Example-Driven BSP and Variational Autoencoderscited by 11
- 2021Dungeon and Platformer Level Blending and Generation using Conditional VAEscited by 10
- 2018Addressing the Fundamental Tension of PCGML with Discriminative Learningcited by 8
Built on by
Citing papers per year: 2023: 9; 2024: 27; 2025: 19; 2026: 13.
68 papers.
Sort:
By area:
- 2024Genie: Generative Interactive Environmentscited by 747
- 2024Large Language Models and Games: A Survey and Roadmapcited by 93
- 2023LLMatic: Neural Architecture Search via Large Language Models and Quality Diversity Optimizationcited by 87
- 2023MarioGPT: Open-Ended Text2Level Generation through Large Language Modelscited by 64
- 2024EnvGen: Generating and Adapting Environments via LLMs for Training Embodied Agentscited by 46
- 2023Beyond Size: How Gradients Shape Pruning Decisions in Large Language Modelscited by 30
- 2024Game Generation via Large Language Modelscited by 25
- 2024DataEnvGym: Data Generation Agents in Teacher Environments with Student Feedbackcited by 25
- 2024GAVEL: Generating Games Via Evolution and Language Modelscited by 24
- 2024Large Language Models and Video Games: A Preliminary Scoping Reviewcited by 23
- 2023Practical PCG Through Large Language Modelscited by 21
- 2024Automating the Search for Artificial Life with Foundation Modelscited by 20
38 more
- 2024DreamCraft: Text-Guided Generation of Functional 3D Environments in Minecraftcited by 17
- 2024Player-Driven Emergence in LLM-Driven Game Narrativecited by 16
- 2026MultiGen: Level-Design for Editable Multiplayer Worlds in Diffusion Game Enginescited by 16
- 2023Games for Artificial Intelligence Research: A Review and Perspectivescited by 14
- 2023The Five-Dollar Model: Generating Game Maps and Sprites from Sentence Embeddingscited by 14
- 2024Unbounded: A Generative Infinite Game of Character Life Simulationcited by 14
- 2024The Ink Splotch Effect: A Case Study on ChatGPT as a Co-Creative Game Designercited by 12
- 2026AI Gamestore: Scalable, Open-Ended Evaluation of Machine General Intelligence with Human Gamescited by 12
- 2024GPT for Games: A Scoping Review (2020-2023)cited by 10
- 2024FactorSim: Generative Simulation via Factorized Representationcited by 10
- 2024DreamGarden: A Designer Assistant for Growing Games from a Single Promptcited by 10
- 2024GPT for Games: An Updated Scoping Review (2020-2024)cited by 10
- 2024Missed Connections: Lateral Thinking Puzzles for Large Language Modelscited by 9
- 2023ChatGPT4PCG Competition: Character-like Level Generation for Science Birdscited by 8
- 2024From Text to Life: On the Reciprocal Relationship between Artificial Life and Large Language Modelscited by 8
- 2024GameTraversalBenchmark: Evaluating Planning Abilities Of Large Language Models Through Traversing 2D Game Mapscited by 8
- 2024Codenames as a Benchmark for Large Language Modelscited by 8
- 2023Testing the Depth of ChatGPT's Comprehension via Cross-Modal Tasks Based on ASCII-Art: GPT3.5's Abilities in Regard to Recognizing and Generating ASCII-Art Are Not Totally Lackingcited by 7
- 2024A Survey on Human-AI Collaboration with Large Foundation Modelscited by 7
- 2024Grammar-based Game Description Generation using Large Language Modelscited by 7
- 2025AnimeGamer: Infinite Anime Life Simulation with Next Game State Predictioncited by 7
- 2024Measuring Diversity of Game Scenarioscited by 5
- 2024ChatPCG: Large Language Model-Driven Reward Design for Procedural Content Generationcited by 5
- 2025Text Speaks Louder than Vision: ASCII Art Reveals Textual Biases in Vision-Language Modelscited by 5
- 2025MARCO: Multi-Agent Code Optimization with Real-Time Knowledge Integration for High-Performance Computingcited by 5
- 2025Divide-Fuse-Conquer: Eliciting "Aha Moments" in Multi-Scenario Gamescited by 4
- 2026Mage: Multi-Axis Evaluation of LLM-Generated Executable Game Scenes Beyond Compile-Pass Ratecited by 4
- 2024Generative Design through Quality-Diversity Data Synthesis and Language Modelscited by 3
- 20243D Building Generation in Minecraft via Large Language Modelscited by 3
- 2025Cardiverse: Harnessing LLMs for Novel Card Game Prototypingcited by 3
- 2025PCGRLLM: Large Language Model-Driven Reward Design for Procedural Content Generation Reinforcement Learningcited by 3
- 2025Word2Minecraft: Generating 3D Game Levels through Large Language Modelscited by 3
- 2025All Stories Are One Story: Emotional Arc Guided Procedural Game Level Generationcited by 3
- 2025Multi-Objective Instruction-Aware Representation Learning in Procedural Content Generation RLcited by 3
- 2026AI for Games in the Foundation Model Eracited by 3
- 2025Grammar and Gameplay-aligned RL for Game Description Generation with LLMscited by 2
- 2025ScriptDoctor: Automatic Generation of PuzzleScript Games via Large Language Models and Tree Searchcited by 2
- 2025Boardwalk: Towards a Framework for Creating Board Games with LLMscited by 2
50 most cited of 68 All 68 (JSON)
- 2026AI for Games in the Foundation Model Era15 Sep 2026
- 2026Programmable Cellular Automata5 Sep 2026
- 2026Procedural Content Metageneration via Program Search and Continual Abstraction Discovery18 Aug 2026
- 2026MAGIC: Transition-Aware Generation of Navigable Multi-Scene Game Worlds with Large Language Models13 Jul 2026
- 2026Knowledge-Conditioned, Single-Pass LLM Synthesis of Executable Unity Game Scenes: A Compiler Error Census across 26 Goal Playable Concepts11 Jul 2026
- 2026AI Native Games: A Survey and Roadmap1 Jul 2026
- 2026Learning Local Constraints for Reinforcement-Learned Content Generators13 May 2026
- 2026Mage: Multi-Axis Evaluation of LLM-Generated Executable Game Scenes Beyond Compile-Pass Rate8 May 2026
- 2026FactorSmith: Agentic Simulation Generation via Markov Decision Process Decomposition with Planner-Designer-Critic Refinement15 Mar 2026
- 2026Bridging Pedagogy and Play: Introducing a Language Mapping Interface for Human-AI Co-Creation in Educational Game Design4 Mar 2026
- 2026MultiGen: Level-Design for Editable Multiplayer Worlds in Diffusion Game Engines3 Mar 2026
- 2026Under the Influence: Quantifying Persuasion and Vigilance in Large Language Models24 Feb 2026
38 more
- 2026AI Gamestore: Scalable, Open-Ended Evaluation of Machine General Intelligence with Human Games19 Feb 2026
- 2025Mortar: Evolving Mechanics for Automatic Game Design31 Dec 2025
- 2025Zero-shot 3D Map Generation with LLM Agents: A Dual-Agent Architecture for Procedural Content Generation11 Dec 2025
- 2025A Markovian Framing of WaveFunctionCollapse for Procedurally Generating Aesthetically Complex Environments12 Sep 2025
- 2025Rule Synergy Analysis using LLMs: State of the Art and Implications27 Aug 2025
- 2025Boardwalk: Towards a Framework for Creating Board Games with LLMs22 Aug 2025
- 2025Multi-Objective Instruction-Aware Representation Learning in Procedural Content Generation RL8 Aug 2025
- 2025All Stories Are One Story: Emotional Arc Guided Procedural Game Level Generation4 Aug 2025
- 2025Fly, Fail, Fix: Iterative Game Repair with Reinforcement Learning and Large Multimodal Models16 Jul 2025
- 2025ScriptDoctor: Automatic Generation of PuzzleScript Games via Large Language Models and Tree Search6 Jun 2025
- 2025Divide-Fuse-Conquer: Eliciting "Aha Moments" in Multi-Scenario Games22 May 2025
- 2025MARCO: Multi-Agent Code Optimization with Real-Time Knowledge Integration for High-Performance Computing6 May 2025
- 2025Analysis of Robustness of a Large Game Corpus4 Apr 2025
- 2025Text Speaks Louder than Vision: ASCII Art Reveals Textual Biases in Vision-Language Models2 Apr 2025
- 2025AnimeGamer: Infinite Anime Life Simulation with Next Game State Prediction1 Apr 2025
- 2025Grammar and Gameplay-aligned RL for Game Description Generation with LLMs20 Mar 2025
- 2025Word2Minecraft: Generating 3D Game Levels through Large Language Models18 Mar 2025
- 2025Hybrid Voting-Based Task Assignment in Role-Playing Games25 Feb 2025
- 2025PCGRLLM: Large Language Model-Driven Reward Design for Procedural Content Generation Reinforcement Learning15 Feb 2025
- 2025Cardiverse: Harnessing LLMs for Novel Card Game Prototyping10 Feb 2025
- 2024Automating the Search for Artificial Life with Foundation Models23 Dec 2024
- 2024Codenames as a Benchmark for Large Language Models16 Dec 2024
- 2024GPT for Games: An Updated Scoping Review (2020-2024)1 Nov 2024
- 2024Unbounded: A Generative Infinite Game of Character Life Simulation24 Oct 2024
- 2024GameTraversalBenchmark: Evaluating Planning Abilities Of Large Language Models Through Traversing 2D Game Maps10 Oct 2024
- 2024DataEnvGym: Data Generation Agents in Teacher Environments with Student Feedback8 Oct 2024
- 2024DreamGarden: A Designer Assistant for Growing Games from a Single Prompt2 Oct 2024
- 2024FactorSim: Generative Simulation via Factorized Representation26 Sep 2024
- 2024Grammar-based Game Description Generation using Large Language Models24 Jul 2024
- 2024GAVEL: Generating Games Via Evolution and Language Models12 Jul 2024
- 2024From Text to Life: On the Reciprocal Relationship between Artificial Life and Large Language Models14 Jun 2024
- 20243D Building Generation in Minecraft via Large Language Models13 Jun 2024
- 2024ChatPCG: Large Language Model-Driven Reward Design for Procedural Content Generation7 Jun 2024
- 2024Generative Design through Quality-Diversity Data Synthesis and Language Models16 May 2024
- 2024GPT for Games: A Scoping Review (2020-2023)27 Apr 2024
- 2024Player-Driven Emergence in LLM-Driven Game Narrative25 Apr 2024
- 2024DreamCraft: Text-Guided Generation of Functional 3D Environments in Minecraft23 Apr 2024
- 2024Missed Connections: Lateral Thinking Puzzles for Large Language Models17 Apr 2024
50 newest of 68 All 68 (JSON)
- 2024Missed Connections: Lateral Thinking Puzzles for Large Language Modelscited by 9
- 2024Codenames as a Benchmark for Large Language Modelscited by 8
- 2025Divide-Fuse-Conquer: Eliciting "Aha Moments" in Multi-Scenario Gamescited by 4
- 2024Zero-Shot Reasoning: Personalized Content Generation Without the Cold Start Problemcited by 1
- 2025Rule Synergy Analysis using LLMs: State of the Art and Implicationscited by 1
- 2026FactorSmith: Agentic Simulation Generation via Markov Decision Process Decomposition with Planner-Designer-Critic Refinementcited by 0
- 2026FactorSmith: Agentic Simulation Generation via Markov Decision Process Decomposition with Planner-Designer-Critic Refinement15 Mar 2026
- 2025Rule Synergy Analysis using LLMs: State of the Art and Implications27 Aug 2025
- 2025Divide-Fuse-Conquer: Eliciting "Aha Moments" in Multi-Scenario Games22 May 2025
- 2024Codenames as a Benchmark for Large Language Models16 Dec 2024
- 2024Missed Connections: Lateral Thinking Puzzles for Large Language Models17 Apr 2024
- 2024Zero-Shot Reasoning: Personalized Content Generation Without the Cold Start Problem15 Feb 2024
- 2024EnvGen: Generating and Adapting Environments via LLMs for Training Embodied Agentscited by 46
- 2024DataEnvGym: Data Generation Agents in Teacher Environments with Student Feedbackcited by 25
- 2025MARCO: Multi-Agent Code Optimization with Real-Time Knowledge Integration for High-Performance Computingcited by 5
- 2025Zero-shot 3D Map Generation with LLM Agents: A Dual-Agent Architecture for Procedural Content Generationcited by 0
- 2025Zero-shot 3D Map Generation with LLM Agents: A Dual-Agent Architecture for Procedural Content Generation11 Dec 2025
- 2025MARCO: Multi-Agent Code Optimization with Real-Time Knowledge Integration for High-Performance Computing6 May 2025
- 2024DataEnvGym: Data Generation Agents in Teacher Environments with Student Feedback8 Oct 2024
- 2024EnvGen: Generating and Adapting Environments via LLMs for Training Embodied Agents18 Mar 2024
- 2024Genie: Generative Interactive Environmentscited by 747
- 2026MultiGen: Level-Design for Editable Multiplayer Worlds in Diffusion Game Enginescited by 16
- 2024Unbounded: A Generative Infinite Game of Character Life Simulationcited by 14
- 2026AI for Games in the Foundation Model Eracited by 3
- 2026AI for Games in the Foundation Model Era15 Sep 2026
- 2026MultiGen: Level-Design for Editable Multiplayer Worlds in Diffusion Game Engines3 Mar 2026
- 2024Unbounded: A Generative Infinite Game of Character Life Simulation24 Oct 2024
- 2024Genie: Generative Interactive Environments23 Feb 2024
- 2023PCGPT: Procedural Content Generation via Transformerscited by 1
- 2023PCGPT: Procedural Content Generation via Transformers3 Oct 2023
- 2024Large Language Models and Games: A Survey and Roadmapcited by 93
- 2023LLMatic: Neural Architecture Search via Large Language Models and Quality Diversity Optimizationcited by 87
- 2023MarioGPT: Open-Ended Text2Level Generation through Large Language Modelscited by 64
- 2023Beyond Size: How Gradients Shape Pruning Decisions in Large Language Modelscited by 30
- 2024Game Generation via Large Language Modelscited by 25
- 2024GAVEL: Generating Games Via Evolution and Language Modelscited by 24
- 2024Large Language Models and Video Games: A Preliminary Scoping Reviewcited by 23
- 2023Practical PCG Through Large Language Modelscited by 21
- 2024Automating the Search for Artificial Life with Foundation Modelscited by 20
- 2024DreamCraft: Text-Guided Generation of Functional 3D Environments in Minecraftcited by 17
- 2024Player-Driven Emergence in LLM-Driven Game Narrativecited by 16
- 2023Games for Artificial Intelligence Research: A Review and Perspectivescited by 14
12 most cited of 51 All 51 (JSON)
- 2026Programmable Cellular Automata5 Sep 2026
- 2026Procedural Content Metageneration via Program Search and Continual Abstraction Discovery18 Aug 2026
- 2026MAGIC: Transition-Aware Generation of Navigable Multi-Scene Game Worlds with Large Language Models13 Jul 2026
- 2026Knowledge-Conditioned, Single-Pass LLM Synthesis of Executable Unity Game Scenes: A Compiler Error Census across 26 Goal Playable Concepts11 Jul 2026
- 2026AI Native Games: A Survey and Roadmap1 Jul 2026
- 2026Learning Local Constraints for Reinforcement-Learned Content Generators13 May 2026
- 2026Mage: Multi-Axis Evaluation of LLM-Generated Executable Game Scenes Beyond Compile-Pass Rate8 May 2026
- 2026Bridging Pedagogy and Play: Introducing a Language Mapping Interface for Human-AI Co-Creation in Educational Game Design4 Mar 2026
- 2026Under the Influence: Quantifying Persuasion and Vigilance in Large Language Models24 Feb 2026
- 2026AI Gamestore: Scalable, Open-Ended Evaluation of Machine General Intelligence with Human Games19 Feb 2026
- 2025Mortar: Evolving Mechanics for Automatic Game Design31 Dec 2025
- 2025A Markovian Framing of WaveFunctionCollapse for Procedurally Generating Aesthetically Complex Environments12 Sep 2025
12 newest of 51 All 51 (JSON)