Technical Report of RoboSpatial Challenge at CVPR 2026: Selective Reasoning Activation and Reference-Frame Disambiguation for Embodied Spatial Reasoning

Yuxiang Xie, Qi Lv, Jianming Xing, Zijian Hong, Xiang Deng +2 more
6/30/2026
cs.CV

Official code: YuxiangXie2003/RoboSpatialBrain (1 star, Apache-2.0, last commit 2026-05-26).

Abstract

Vision-language models achieve strong general perception but often struggle with the spatial reasoning required for embodied tasks. We present RoboSpatialBrain, our submission to the RoboSpatial Challenge at the Embodied Reasoning in Action Workshop, CVPR 2026, built on RoboBrain2.5-8B-NV. RoboSpatialBrain combines two training-free, inference-time mechanisms: a forced <think> prefix activation strategy paired with a task-specific post-prompt that elicits deliberate reasoning on context and compatibility tasks, and an explicit reference-frame redirection pipeline that resolves camera-centric and object-centric ambiguity for context tasks. We additionally explore fine-tuning RoboBrain2.5 on compatibility data and present a detailed analysis of its interaction with prompting. RoboSpatialBrain achieved first place in the RoboSpatial Challenge, with an overall success rate of 80.9\% on RoboSpatial-Home. Code is available at https://github.com/YuxiangXie2003/RoboSpatialBrain.

View on arXivView PDF

Code Implementations(1)

YuxiangXie2003/RoboSpatialBrainOfficial100%
Apache-2.0
10PythonMay 22, 20264 months agoApache-2.0

Similar Papers

EPIC-Bench: A Perception-Centric Benchmark for Fine-Grained Embodied Visual Grounding in Vision-Language Models

86% match2026

From Spatial to Actions: Grounding Vision-Language-Action Model in Spatial Foundation Priors

86% match2025

MindEdit-Bench: Benchmarking Object-Level Counterfactual Spatial Reasoning in VLMs from In-the-Wild Photos

85% match2026

Plan-Bench: Can Existing VLMs Go Well Beyond Perception to Reason and Plan in Embodied Worlds?

85% match1 repo2026

Embodied4C: Measuring What Matters for Embodied Vision-Language Navigation

85% match2025

ESPIRE: A Diagnostic Benchmark for Embodied Spatial Reasoning of Vision-Language Models

85% match2 repos122026

Cite this paper

@article{xie2026technical,
  title  = {Technical Report of RoboSpatial Challenge at CVPR 2026: Selective Reasoning Activation and Reference-Frame Disambiguation for Embodied Spatial Reasoning},
  author = {Yuxiang Xie and Qi Lv and Jianming Xing and Zijian Hong and Xiang Deng and Weili Guan and Liqiang Nie},
  year   = {2026},
  eprint = {2606.31645},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2606.31645v1}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2606.31645.svg)](https://codeofpaper.com/paper/2606.31645)

Discussion