VibeWorlding:多模态智能体能否端到端构建3D开放世界?
TLDR
提出VibeWorlding基准与RL训练框架,用于多模态智能体从用户查询构建3D开放世界;当前MLLMs成功率低于60%。
评分理由
The paper introduces a substantial benchmark and training environment for multimodal 3D world construction, with empirical evaluation of frontier models. However, its core contribution is agent benchmarking and RL post-training rather than world modeling or dynamics prediction, so world-model-related keywords are only weakly relevant.
Read-first 评分解释
综合优先阅读分 24.6,由主题、引用、图谱、方法、可复现性和近期性等信号加权得到。 原始总分保留为 11。
研究版图角色
前沿论文
排序敏感性
稳定性:volatile;排名波动范围:46。