Research
Chain-of-Thought Degrades Visual Spatial Reasoning Capabilities of Multimodal LLMs
arXiv:2604.16060v1 Announce Type: cross Abstract: Multimodal Reasoning Models (MRMs) leveraging Chain-of-Thought (CoT) based thinking have revolutionized mathematical and logical problem-solving. Howe
arXiv:2604.16060v1 Announce Type: cross Abstract: Multimodal Reasoning Models (MRMs) leveraging Chain-of-Thought (CoT) based thinking have revolutionized mathematical and logical problem-solving. However, we show that this paradigm struggles with generalized spatial intelligence. We perform a comprehensive evaluation of seventeen models across thirteen spatial benchmarks and identify a critical gap: CoT prompting consistently degrades performance in visual spatial reasoning. Furthermore, through a novel No-Image++ ablation, we demonstrate that MRMs and CoT prompted MLMs suffer from severe shortcut learning, and hallucinate visual details from textual priors even when the image is absent. These findings challenge the efficacy of text-only CoT for spatial tasks and underscore the need for vision-centric reasoning paradigms.
Related
- Rectifying LLM Thought from Lens of Optimization
- Faithful-First Reasoning, Planning, and Acting for Multimodal LLMs
- Critical-CoT: A Robust Defense Framework against Reasoning-Level Backdoor Attacks in Large Language Models
- Revisiting the Capacity Gap in Chain-of-Thought Distillation from a Practical Perspective
Source: arXiv cs.AI | 2026-04-20