一句话：提出解耦式 VLM 智能体框架 AstraGame，通过并行感知推理、语义缓存与动作协议，显著提升大规模游戏测试的效率与效果。
论文针对自动化游戏测试中直接端到端使用视觉语言模型（VLM）所带来的执行时间开销大、token 成本高且测试效果有限的问题，提出工业级框架 AstraGame。其核心是解耦式架构：让专用小模型与大型认知 VLM 协同工作，具体包括并行化感知与推理任务以实现时间饱和、引入组件级（widget-level）语义缓存把重复推理转化为高效缓存查找，以及结构化动作协议以降低 token 消耗。评估显示，相比当前实践方案，探索覆盖率提升 37.78%；在微信 24,000 个独特小游戏上的大规模部署中，端到端延迟降低 58%，产出有效且完整内容审核素材的游戏比例从 38% 提升到 63%，并已发现超过 180,000 个可供开发者处理的功能与合规问题，说明工业级智能需要从孤立的 VLM 推理转向整体系统编排。
Automated game testing requires deep semantic reasoning to navigate complex dynamic environments, but directly applying end-to-end Vision-Language Models (VLMs) at an industrial scale incurs prohibitive execution-time overhead and token costs with limited testing effectiveness. To resolve the trade-off between reasoning depth and execution speed in automated game testing, we present AstraGame, a VLM-based game testing framework that leverages a decoupled architecture to orchestrate collaboration between specialized small models and large cognitive VLMs. Specifically, we propose to parallelize perception and reasoning tasks for temporal saturation, implement widget-level semantic caching to transform repetitive inference into efficient cache lookup, and structure action protocols for token efficiency. Evaluation results show that AstraGame substantially improves exploration coverage by 37.78% compared to state-of-the-practice approaches. Extensive deployment across 24,000 unique mini-games on WeChat, one of the world's largest mini-game platforms, demonstrates that AstraGame achieves a 58% reduction in end-to-end latency, increasing the proportion of games yielding valid and comprehensive materials for game content moderation from 38% to 63%. To date, AstraGame has identified over 180,000 functional and compliance issues actionable by developers, validating that industrial-scale intelligence requires shifting focus from isolated VLM inference to comprehensive system orchestration.
@inproceedings{GuoLWXDRYX26,
title = {AstraGame: Effective and Efficient VLM Agent Serving for Large-Scale Game Testing in an Industry Setting},
author = {Yuzhe Guo and Haochuan Lu and Mengzhou Wu and Ting Xiong and Yuetang Deng and Dezhi Ran and Wei Yang and Tao Xie},
booktitle = {FSE},
year = {2026},
doi = {10.1145/3803437.3805252}
}