| title | SpatialOPSD: Self-Distilling Spatial Intelligence from Verified Coding Agent Traces |
|---|---|
| authors | Rongxue Li, Meng Yang, Yiru Mao, Yongliang Tao, Lulu Hu, Bin Yang, Zhao Xu, Weihua Luo, Bowen Xu |
| date | 2026-10-08 |
| venue | arXiv (preprint, not peer reviewed) |
| id | arXiv:2610.11366 |
| affiliation | ATH-MaaS |
| upvotes | 15 (Hugging Face Daily Papers, 2026-11-01) |
| links | abs pdf hf |
| abstract | Spatial coding agents significantly improve spatial reasoning in Multimodal Large Language Models (MLLMs) by using external tools to generate verified execution traces. However, this paradigm inherently suffers from prohibitive inference-time overhead and external dependencies. In this paper, we explore whether an MLLM can internalize this agentic capability to operate entirely tool-free. We begin with a simple observation: prompting an MLLM with summarized execution traces of a spatial coding agent naturally unlocks the model's internal spatial Chain-of-Thought (CoT). Motivated by this, we introduce SpatialOPSD, an on-policy self-distillation framework that internalizes spatial reasoning into a standalone MLLM by formulating verified agent traces as privileged information. To mitigate privileged-information leakage during distillation, we introduce Repetition-Aware Distillation, which combines repetition masking with unlikelihood regularization. Experiments across multiple benchmarks demonstrate that self-distilling SpatialOPSD achieves higher average accuracy than SFT and GRPO on both spatial and OOD datasets, exhibiting superior performance and generalization. |
cite
@misc{li2026spatialopsd,
title={SpatialOPSD: Self-Distilling Spatial Intelligence from Verified Coding Agent Traces},
author={Rongxue Li and Meng Yang and Yiru Mao and Yongliang Tao and Lulu Hu and Bin Yang and Zhao Xu and Weihua Luo and Bowen Xu},
year={2026},
eprint={2610.11366},
archivePrefix={arXiv},
url={https://arxiv.org/abs/2610.11366}
}