Recursive Game Creator: An Agentic Product-Level Experience-Oriented Game Harness
Abstract
Recent game design agents have made substantial progress in generating playable games. However, program correctness does not ensure an enjoyable experience for players. We present Recursive Game Creator, an experience-oriented harness to advance agentic game development from rough game prototypes into entertaining games. Recursive Game Creator organizes recursive development around four components: Designer, Builder, Player, and Reviewer. The Designer translates user instructions and Reviewer's feedback into detailed plans. The Builder turns these plans into candidate games. The coding-native Player creates and executes reusable policies through programmatic interfaces to efficiently collect diverse gameplay trajectories, mitigating evaluation bias caused by slow GUI-based collection. The Reviewer uses carefully designed trajectory-based metrics to induce player preferences, integrating with visual evidence and explicit textual preferences to evaluate games against game-specific criteria. Finally, the Reviewer accepts the better version and provides improvement reviews for the next round, closing the recursive loop. Our method achieves state-of-the-art overall performance of 77.89 on GameCraft-Bench. On GameASG-Bench, it achieves a strict task success rate of 53.2%, a 34.1% improvement over the same-model baseline, and the highest mean runtime-check pass rate at 93.4% among compared methods. A user study shows longer playtime and higher ratings. Code is coming soon.
Community
We introduce Recursive Game Creator, an experience-oriented agentic framework that transforms playable game prototypes into engaging games through recursive development.
Inspired by real game studios, our framework brings together four collaborative roles: Designer, Builder, Player, and Reviewer. A key innovation is our Coding-Native Player, which efficiently collects diverse gameplay trajectories through programmatic interaction, enabling more comprehensive and less biased game evaluation.
Our framework achieves state-of-the-art performance on GameCraft-Bench (77.89), a 53.2% strict task success rate on GameASG-Bench, and a 93.4% runtime-check pass rate. User studies further demonstrate improvements in player engagement and game quality.
We hope this work provides a step toward AI agents that can create games people genuinely enjoy playing!
This is an automated message from the Librarian Bot. I found the following papers similar to this paper.
The following papers were recommended by the Semantic Scholar API
- SWE-Game: Can Coding Agents Build the Games We Want? (2026)
- GameXpert-Bench: How Far Are Coding Agents from Expert Game Development? (2026)
- RSIGame: Autonomous Agentic Game Development with Recursive Self-improvement (2026)
- GameASG-Bench: Benchmarking Autonomous Software Generation for Game Development (2026)
- GAMEGO: Training Game-Dev Agents with Synthetic Trajectories Anchored in Real-World Assets (2026)
- GameLogicBench: Evaluating Coding Agents on Runtime Game Logic with Tick-Level State Assertions (2026)
- CraftBench-UE: Deterministic Evaluation for Coding Agents in Unreal Engine (2026)
Please give a thumbs up to this comment if you found it helpful!
If you want recommendations for any Paper on Hugging Face checkout this Space
You can directly ask Librarian Bot for paper recommendations by tagging it in a comment: @librarian-bot recommend
Get this paper in your agent:
hf papers read 2610.08621 Don't have the latest CLI?
curl -LsSf https://hf.co/cli/install.sh | bash Models citing this paper 0
No model linking this paper
Datasets citing this paper 0
No dataset linking this paper