以多模态基础模型扩展空间智能
Scaling Spatial Intelligence with Multimodal Foundation Models
该工作系统构建了800万空间能力数据并微调多模态基础模型，为视觉研究者提供了可复用的空间智能训练数据构建范式与评测基准，同时揭示了数据缩放对空间推理泛化能力的影响。
原文依据 we explore scaling up multimodal foundation models to cultivate spatial intelligence within the SenseNova-SI family
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论尽管取得了显著进展，多模态基础模型在空间智能方面仍表现出令人惊讶的缺陷。在本工作中，我们探索扩展多模态基础模型以培养空间智能，构建了 SenseNova-SI 系列，基于已有的多模态基础，包括视觉理解模型（即 Qwen3-VL 和 InternVL3）以及统一理解与生成模型（即 Bagel）。我们采取原则性方法构建高性能且鲁棒的空间智能，通过系统性地整理 SenseNova-SI-8M：在严格的空间能力分类体系下的八百万个多样化数据样本。SenseNova-SI 在广泛的空间智能基准上展现了前所未有的性能：VSI-Bench 上 68.8%，MMSI 上 43.3%，MindCube 上 85.7%，ViewSpatial 上 54.7%，SITE 上 47.7%，BLINK 上 63.9%，3DSR 上 55.5%，EmbSpatial 上 72.0%，同时保持强大的通用多模态理解能力（例如 MMBench-En 上 84.9%）。更重要的是，我们分析了数据缩放的影响，讨论了多样化数据训练带来的涌现泛化能力的早期迹象，分析了过拟合和语言捷径的风险，提出了关于空间思维链推理的初步研究，并验证了潜在的下游应用。所有新训练的多模态基础模型均已公开释放。
英文原文摘要 1391 字符
Despite remarkable progress, multimodal foundation models still exhibit surprising deficiencies in spatial intelligence. In this work, we explore scaling up multimodal foundation models to cultivate spatial intelligence within the SenseNova-SI family, built upon established multimodal foundations including visual understanding models (i.e., Qwen3-VL and InternVL3) and unified understanding and generation models (i.e., Bagel). We take a principled approach to constructing high-performing and robust spatial intelligence by systematically curating SenseNova-SI-8M: eight million diverse data samples under a rigorous taxonomy of spatial capabilities. SenseNova-SI demonstrates unprecedented performance across a broad range of spatial intelligence benchmarks: 68.8% on VSI-Bench, 43.3% on MMSI, 85.7% on MindCube, 54.7% on ViewSpatial, 47.7% on SITE, 63.9% on BLINK, 55.5% on 3DSR, and 72.0% on EmbSpatial, while maintaining strong general multimodal understanding (e.g., 84.9% on MMBench-En). More importantly, we analyze the impact of data scaling, discuss early signs of emergent generalization capabilities enabled by diverse data training, analyze the risk of overfitting and language shortcuts, present a preliminary study on spatial chain-of-thought reasoning, and validate the potential downstream application. All newly trained multimodal foundation models are publicly released.