RSIGame: Autonomous Agentic Game Development with Recursive Self-improvement
cs.CL, cs.GT, cs.LG, cs.MA
Submitted: 2026-09-30
Updated: 2026-09-30
Code: https://github.com/WenyiWU0111/RSIGame2https:
License: http://creativecommons.org/licenses/by/4.0/
The gist: Recent advances in large language models have made automatic game generation increasingly feasible, yet reliably improving generated games beyond a playable version remains challenging.
Terminology
Abstract
Recent advances in large language models have made automatic game generation increasingly feasible, yet reliably improving generated games beyond a playable version remains challenging. Naive iterative refinement can easily overfit a small set of test cases, producing fragile games with unresolved bugs, missing behaviors, and poor generalization to broader player interactions. We introduce RSIGame, an autonomous agentic game development framework with recursive self-improvement. RSIGame organizes development into complementary local and global loops. Concretely, a local explore-diagnose-improve loop broadly explores the executable game, diagnoses and prioritizes discovered issues, and performs evidence-grounded revision, where an evolving checklist continually accumulates new testing and improvement guidance. A global loop tracks overall quality, preserves the best checkpoint, and detects saturation or regression over long-horizon development. Beyond test-time improvement, RSIGame further internalizes successful development experience into the generator through training. Across 140 GameCraft-Bench tasks, two game engines, and five generators, RSIGame consistently improves game quality under matched development budgets. Notably, experience internalization enables Qwen3.8-27B to reach 61.38 on Godot and 58.53 on Phaser, exceeding GPT-5.5 one-shot scores while reducing Qwen's generation tokens by 11 times.
Sources
- Teaching Large Language Models to Self-Debug
- GameDevBench: Evaluating Agentic Capabilities Through Game Development
- QLoRA: Efficient Finetuning of Quantized LLMs
- MetaGPT: Meta Programming for A Multi-Agent Collaborative Framework
- LoRA: Low-Rank Adaptation of Large Language Models
- Decoupled Weight Decay Regularization
- Self-Refine: Iterative Refinement with Self-Feedback
- Reflexion: Language Agents with Verbal Reinforcement Learning
- Co-Evolving LLM Coder and Unit Tester via Reinforcement Learning
- Meta-Rewarding Language Models: Self-Improving Alignment with LLM-as-a-Meta-Judge
- SWE-agent: Agent-Computer Interfaces Enable Automated Software Engineering
- ReAct: Synergizing Reasoning and Acting in Language Models
- STaR: Bootstrapping Reasoning With Reasoning
- Judging LLM-as-a-Judge with MT-Bench and Chatbot Arena
- RSIAgent: Autonomous Exploration for Recursive Self-improvement in New Environments
Related papers
- Exploring Solution Divergence and Its Effect on Large Language Model Problem Solving
- Ishigaki-IDS-Bench: A Benchmark for Generating Information Delivery Specification from BIM Information Requirements
- Subliminal Steering: Stronger Encoding of Hidden Signals
- MedStruct-S: A Benchmark for Key Discovery, Key-Conditioned QA and Semi-Structured Extraction from OCR Clinical Reports
- The End of Transformers? On Challenging Attention and the Rise of Sub-Quadratic Architectures
- Untangling the Mechanisms of Misleading Context in Medical Question Answering