Current Large Multimodal Models (LMMs) struggle with spatial reasoning tasks requiring viewpoint-dependent understanding, largely because they are confined to a single, static observation. We propose Thinking with Novel Views (TwNV), a paradigm that integrates generative novel-view synthesis into the reasoning loop: a Reasoner LMM identifies spatial ambiguity, instructs a Painter to synthesize an alternative viewpoint, and re-examines the scene with the additional evidence. Through systematic experiments we address three research questions. (1) Instruction format: numerical camera-pose specifications yield more reliable view control than free-form language. (2) Generation fidelity: synthesized view quality is tightly coupled with downstream spatial accuracy. (3) Inference-time visual scaling: iterative multi-turn view refinement further improves performance, echoing recent scaling trends in language reasoning. Across four spatial subtask categories and four LMM architectures (both closed- and open-source), TwNV consistently improves accuracy by +1.3 to +3.9 pp, with the largest gains on viewpoint-sensitive subtasks. These results establish novel-view generation as a practical lever for advancing spatial intelligence of LMMs.
@article{arxiv.2605.10588,
title = {Thinking with Novel Views: A Systematic Analysis of Generative-Augmented Spatial Intelligence},
author = {Yanbing Zhang and Bo Wang and Jianhui Liu and Nan Jiang and Jiaxiu Jiang and Haoze Sun and Yijun Yang and Shenghe Zheng and Lin Song and Haoyang Huang and Nan Duan and Wenbo Li},
journal= {arXiv preprint arXiv:2605.10588},
year = {2026}
}