Which Pretraining Paradigm Better Serves Spatial Intelligence? An Empirical Comparison of Vision-Language and Video Generation Models

• URL: http://arxiv.org/abs/2605.28132v1
• Date: Wed, 27 May 2026 08:20:04 GMT
• Title: Which Pretraining Paradigm Better Serves Spatial Intelligence? An Empirical Comparison of Vision-Language and Video Generation Models
• Abstract summary: We present the first systematic frozen-feature probing study of Vision-Language Models (VLMs) and Video Generation Models (VGMs)<n>Using the lightweight probe, our framework enables a controlled comparison of what information is already encoded in frozen representations from two model families.<n>Results reveal a clear complementarity: VLMs are stronger at semantic tagging and instance grouping, while VGMs provide more accessible signals for dense geometry and camera motion.
• Score: 31.851717131965824
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Spatial intelligence requires visual representations that capture both semantic objects and geometric structure in the physical world. To support this, two major pre-training schemes are now widely used as foundation backbones: Vision-Language Models (VLMs), which use language supervision to align visual observations with semantic concepts, and Video Generation Models (VGMs), which learn from temporally evolving visual worlds. However, it still remains unclear which pre-training scheme provides a better representation substrate for spatial intelligence. In this paper, we present the first systematic frozen-feature probing study of VLMs and VGMs across three representative axes of spatial intelligence: semantic tagging, instance grouping, and 3D geometry prediction. Using the lightweight probe, our framework enables a controlled comparison of what information is already encoded in frozen representations from two model families. Experimental results reveal a clear complementarity: VLMs are stronger at semantic tagging and instance grouping, while VGMs provide more accessible signals for dense geometry and camera motion. Moreover, a naive fusion of the two already yields a representation that excels at both geometry and semantics, suggesting a promising direction for building stronger spatial-intelligence backbones by effectively integrating features from both model families. Our code is available at \href{https://github.com/om-ai-lab/Probing-VLM-VGM}{https://github.com/om-ai-lab/Probing-VLM-VGM}.

Related papers

• Learning Geometric Representations from Videos for Spatial Intelligent Multimodal Large Language Models [32.477423778941386]
  We present GeoVR, a framework that learns geometric representations using purely 2D video sequences.<n>Rather than employing superficial feature mixing, GeoVR reshapes the internal representations of the MLLM by distilling geometry knowledge.<n>Experiments on spatial reasoning benchmarks demonstrate that GeoVR achieves state-of-the-art performance.
  arXiv  Detail & Related papers  (2026-06-04T08:11:12Z)
• Dual-Pathway Geometry-Aware MLLM for Spatial Intelligence [63.11502936651325]
  We propose GAMSI, a dual-pathway Geometry-Aware MLLM for Spatial Intelligence.<n>It takes only RGB images as input while internalizing both forms of geometric prior within a unified autoregressive backbone.<n>GAMSI achieves state-of-the-art performance on seven spatial intelligence benchmarks.
  arXiv  Detail & Related papers  (2026-05-25T01:33:19Z)
• GeoWorld-VLM: Geometry from World Models for Vision-Language Models [10.86505613923278]
  Modern Vision-Language Models (VLMs) achieve strong semantic recognition, yet remain brittle on elementary spatial relations.<n>We introduce GeoWorld-VLM, a VLM-side distillation framework that transfers geometric structure from frozen video world models intoVLMs.<n>GeoWorld-VLM fine-tunes only the image encoder and multimodal projector, aligning post-projector image features with intermediate world-model representations while leaving the main backbone frozen.
  arXiv  Detail & Related papers  (2026-05-15T23:52:11Z)
• VEGA: Visual Encoder Grounding Alignment for Spatially-Aware Vision-Language-Action Models [55.12929235609365]
  Visual backbones of current vision-language-action (VLA) models are predominantly pretrained on 2D image data without explicit 3D geometric supervision.<n>Existing implicit grounding methods partially address this by aligning VLA features with those of 3D-aware foundation models.<n>We propose VEGA, a framework that directly aligns the output of the VLA's visual encoder with spatially-aware features from DINOv2-FiT3D.
  arXiv  Detail & Related papers  (2026-05-11T12:44:26Z)
• Proxy3D: Efficient 3D Representations for Vision-Language Models via Semantic Clustering and Alignment [50.02189698630855]
  We propose a Proxy3D method with compact yet comprehensive 3D proxy representations for the vision modality.<n>Our method achieves competitive or state-of-the-art performance in 3D visual question answering, visual grounding and general spatial intelligence benchmarks.
  arXiv  Detail & Related papers  (2026-05-08T17:50:47Z)
• HERMES++: Toward a Unified Driving World Model for 3D Scene Understanding and Generation [83.31948299340782]
  HERMES++ is a unified driving world model that integrates 3D scene understanding and future geometry prediction within a single framework.<n>Hermes++ achieves strong performance, outperforming specialist approaches in both future point cloud prediction and 3D scene understanding tasks.
  arXiv  Detail & Related papers  (2026-04-30T17:59:58Z)
• TagaVLM: Topology-Aware Global Action Reasoning for Vision-Language Navigation [70.23578202012048]
  Vision-Language Navigation (VLN) presents a unique challenge for Large Vision-Language Models (VLMs) due to their inherent architectural mismatch.<n>We propose TagaVLM (Topology-Aware Global Action reasoning), an end-to-end framework that explicitly injects topological structures into the VLM backbone.<n>To enhance topological node information, an Interleaved Navigation Prompt strengthens node-level visual-text alignment.<n>With the embedded topological graph, the model is capable of global action reasoning, allowing for robust path correction.
  arXiv  Detail & Related papers  (2026-03-03T13:28:07Z)
• EagleVision: A Dual-Stage Framework with BEV-grounding-based Chain-of-Thought for Spatial Intelligence [10.889641815961133]
  spatial intelligence approaches typically attach 3D cues to 2D reasoning pipelines or MLLMs with black-box reconstruction modules.<n>We present EagleVision, a framework for progressive spatial cognition through macro perception and micro verification.
  arXiv  Detail & Related papers  (2025-12-17T07:51:36Z)
• Evo-0: Vision-Language-Action Model with Implicit Spatial Understanding [11.222744122842023]
  We introduce a plug-and-play module that implicitly incorporates 3D geometry features into Vision-Language-Action (VLA) models.<n>Our method significantly improves the performance of state-of-the-art VLA models across diverse scenarios.
  arXiv  Detail & Related papers  (2025-07-01T04:05:47Z)
• Spatial-MLLM: Boosting MLLM Capabilities in Visual-based Spatial Intelligence [13.168559963356952]
  We present Spatial-MLLM, a novel framework for visual-based spatial reasoning from purely 2D observations.<n>Our key insight is to unleash the strong structure prior to the feed-forward visual geometry foundation model.<n>A connector then integrates both features into unified visual tokens for enhanced spatial understanding.
  arXiv  Detail & Related papers  (2025-05-29T17:59:04Z)
• SimIPU: Simple 2D Image and 3D Point Cloud Unsupervised Pre-Training for Spatial-Aware Visual Representations [85.38562724999898]
  We propose a 2D Image and 3D Point cloud Unsupervised pre-training strategy, called SimIPU. Specifically, we develop a multi-modal contrastive learning framework that consists of an intra-modal spatial perception module and an inter-modal feature interaction module. To the best of our knowledge, this is the first study to explore contrastive learning pre-training strategies for outdoor multi-modal datasets.
  arXiv  Detail & Related papers  (2021-12-09T03:27:00Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.