mlsanityData to 14 Oct 2026

When Depth Is Better Told Than Shown: Depth-Ordinal Prompting for Vision-Language Spatial Reasoning

arXiv 2607.11173 ↗First posted 13 Jul 2026cs.CVcited by 0

On the map: LLM reasoning → Spatial and embodied reasoning

also in 1 more
Abstract (arXiv)

Vision-language models (VLMs) are expected to reason about physical space -- which object is closer, what lies behind what, and how objects are arranged in 3D -- yet they still struggle with such spatial judgments. A natural remedy is to show the model a depth map, but we find that this can make performance worse. We show that depth is not absent: it reaches the language model, but becomes difficult to access for downstream reasoning, while rendered pseudo-depth maps act as noisy auxiliary images that frozen VLMs cannot easily regulate. We propose Depth-Ordinal Prompting (DOP), a training-free method that converts monocular depth into a single question-targeted ordinal text cue at the queried objects, without adding a depth image, training a module, injecting features, or using labels. Our key finding is form dependence: the same depth signal can hurt when shown as an image but help when told as text.Across benchmarks, models, and depth estimators, DOP improves spatial reasoning when pseudo-depth provides reliable object-level ordering and remains largely neutral in strong original-image regimes. It is also competitive with the strongest training-free depth-prompting alternative while being simpler and more targeted.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Spatial and embodied reasoning, one of 17 directions in LLM reasoning. Builds on 28 of its 32 references that are in the map. First posted on arXiv 13 Jul 2026.

Builds on

28 of 32 references.

Closest ancestors · 8

Same direction, newest first

  1. 2026Predicting When to Trust Vision-Language Models for Spatial Reasoning14 Jan
  2. 2025RoboRefer: Towards Spatial Referring with Reasoning in Vision-Language Models for Robotics4 Jun
  3. 2025SSR: Enhancing Depth Perception in Vision-Language Models via Rationale-Guided Spatial Reasoning18 May
  4. 2024Reasoning Paths with Reference Objects Elicit Quantitative Spatial Reasoning in Large Vision-Language Models15 Sep
  5. 2024Coarse Correspondences Boost Spatial-Temporal Reasoning in Multimodal Language Model1 Aug
  6. 2024SpatialPIN: Enhancing Spatial Reasoning Capabilities of Vision-Language Models through Prompting and Interacting 3D Priors18 Mar
  7. 2024SpatialVLM: Endowing Vision-Language Models with Spatial Reasoning Capabilities22 Jan
  8. 2022Visual Spatial Reasoning30 Apr

Other references · 20

  1. 2021Vision Transformers for Dense Predictioncited by 1,803
  2. 2023Evaluating Object Hallucination in Large Vision-Language Modelscited by 1,729
  3. 2023MME: A Comprehensive Evaluation Benchmark for Multimodal Large Language Modelscited by 1,652
  4. 2023Inference-Time Intervention: Eliciting Truthful Answers from a Language Modelcited by 1,302
  5. 2024Cambrian-1: A Fully Open, Vision-Centric Exploration of Multimodal LLMscited by 838
  6. 2023Just Ask for Calibration: Strategies for Eliciting Calibrated Confidence Scores from Language Models Fine-Tuned with Human Feedbackcited by 812
  7. 2018Object Hallucination in Image Captioningcited by 669
  8. 2023ZoeDepth: Zero-shot Transfer by Combining Relative and Metric Depthcited by 657
  9. 2024BLINK: Multimodal Large Language Models Can See but Not Perceivecited by 550
  10. 2023Adaptive Chameleon or Stubborn Sloth: Revealing the Behavior of Large Language Models in Knowledge Conflictscited by 341
  11. 2023Interpreting CLIP's Image Representation via Text-Based Decompositioncited by 196
  12. 2024SpatialBot: Precise Spatial Understanding with Vision Language Modelscited by 152
8 more
  1. 2024EmbSpatial-Bench: Benchmarking Spatial Understanding for Embodied Tasks with Large Vision-Language Modelscited by 107
  2. 2024VL-Uncertainty: Detecting Hallucination in Large Vision-Language Model via Uncertainty Estimationcited by 73
  3. 2024Scaffolding Coordinates to Promote Vision-Language Coordination in Large Multi-Modal Modelscited by 57
  4. 2024Selective "Selective Prediction": Reducing Unnecessary Abstention in Vision-Language Reasoningcited by 40
  5. 2025Adaptive Retrieval Without Self-Knowledge? Bringing Uncertainty Back Homecited by 33
  6. 2025SD-VLM: Spatial Measuring and Understanding with Depth-Encoded Vision-Language Modelscited by 25
  7. 2026VIB-Probe: Detecting and Mitigating Hallucinations in Vision-Language Models via Variational Information Bottleneckcited by 8
  8. 2025ByDeWay: Boost Your multimodal LLM with DEpth prompting in a Training-Free Waycited by 3

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026Seeing Depth but Choosing Cues in Vision-Language Models
  2. 2026Do Vision-Language Models Understand 3D Scenes or Just Catalogue Objects?
  3. 2026Disentangling Pictorial Cue Understanding from Language Bias in VLMs via Depth Ordering Task
  4. 2026Distilling Routed 3D Privilege for Spatial Reasoning in Vision-Language Models
  5. 2025Pursuing Minimal Sufficiency in Spatial Reasoning
  6. 2026DepthEvidence: Unifying Metric Depth Prediction and Geometric Reasoning in Multimodal Language Models
  7. 2026The Spatial Blindspot of Vision-Language Models
  8. 2026Spatial Reasoning is Not a Free Lunch: A Controlled Study on LLaVA