PAI-Bench：面向物理 AI 的综合基准
PAI-Bench: A Comprehensive Benchmark For Physical AI
做视频生成或视频理解的人可以直接用 PAI-Bench 的 2,808 个真实案例和任务对齐指标，检验模型是否保持物理一致的动力学，而不是只看视觉保真度；也可借鉴其条件视频生成与视频理解统一评测的协议设计。
原文依据 a unified and comprehensive benchmark that evaluates perception and prediction capabilities across video generation, conditional video generation, and video understanding
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论物理 AI 旨在开发能够感知并预测真实世界动态的模型；然而，当前多模态大语言模型与视频生成模型在多大程度上支持这些能力，仍缺乏充分认识。我们提出物理 AI 基准（PAI-Bench），这是一个统一且全面的基准，在视频生成、条件视频生成与视频理解三类任务上评估感知与预测能力，包含 2,808 个真实世界案例，并配有与任务对齐、旨在刻画物理合理性与领域特定推理的指标。我们的研究对近期模型进行了系统性评估，结果表明：视频生成模型尽管视觉保真度强，却往往难以维持物理上连贯的动态；而多模态大语言模型在预测与因果解释方面表现有限。这些观察表明，当前系统在处理物理 AI 的感知与预测需求方面仍处于早期阶段。总之，PAI-Bench 为评估物理 AI 建立了贴近现实的基础，并凸显了未来系统必须解决的关键差距。
英文原文摘要 1166 字符
Physical AI aims to develop models that can perceive and predict real-world dynamics; yet, the extent to which current multi-modal large language models and video generative models support these abilities is insufficiently understood. We introduce Physical AI Bench (PAI-Bench), a unified and comprehensive benchmark that evaluates perception and prediction capabilities across video generation, conditional video generation, and video understanding, comprising 2,808 real-world cases with task-aligned metrics designed to capture physical plausibility and domain-specific reasoning. Our study provides a systematic assessment of recent models and shows that video generative models, despite strong visual fidelity, often struggle to maintain physically coherent dynamics, while multi-modal large language models exhibit limited performance in forecasting and causal interpretation. These observations suggest that current systems are still at an early stage in handling the perceptual and predictive demands of Physical AI. In summary, PAI-Bench establishes a realistic foundation for evaluating Physical AI and highlights key gaps that future systems must address.