Spatial intelligence is crucial for vision--language models (VLMs) in the physical world, yet many benchmarks evaluate largely unconstrained scenes where models can exploit 2D shortcuts. We introduce SSI-Bench, a VQA benchmark for spatial reasoning on constrained manifolds, built from complex real-world 3D structures whose feasible configurations are tightly governed by geometric, topological, and physical constraints. SSI-Bench contains 1,000 ranking questions spanning geometric and topological reasoning and requiring a diverse repertoire of compositional spatial operations, such as mental rotation, cross-sectional inference, occlusion reasoning, and force-path reasoning. It is created via a fully human-centered pipeline: ten researchers spent over 400 hours curating images, annotating structural components, and designing questions to minimize pixel-level cues. Evaluating 31 widely used VLMs reveals a large gap to humans: the best open-source model achieves 22.2% accuracy and the strongest closed-source model reaches 33.6%, while humans score 91.6%. Encouraging models to think yields only marginal gains, and error analysis points to failures in structural grounding and constraint-consistent 3D reasoning. Project page: https://ssi-bench.github.io.

翻译：空间智能对于视觉-语言模型（VLMs）在物理世界中的运用至关重要，然而现有基准测试大多评估的是几乎无约束的场景，使得模型可以利用二维捷径。我们提出了SSI-Bench，这是一个用于约束流形上空间推理的视觉问答基准，它基于复杂的真实世界三维结构构建，其可行配置受到几何、拓扑和物理约束的严格限制。SSI-Bench包含1000个排序问题，涵盖几何与拓扑推理，并要求模型具备多样化的组合空间操作能力，例如心理旋转、截面推断、遮挡推理和力路径推理。该基准通过完全以人为中心的流程创建：十名研究人员花费超过400小时来筛选图像、标注结构组件并设计问题，以最小化像素级线索。对31个广泛使用的VLMs进行评估后，发现其与人类表现存在巨大差距：最佳开源模型的准确率为22.2%，最强的闭源模型达到33.6%，而人类得分高达91.6%。鼓励模型进行思考仅带来边际收益，错误分析则指向其在结构基础化和约束一致的三维推理方面的失败。项目页面：https://ssi-bench.github.io。