mlsanityData to 14 Oct 2026

SceneCOT: Eliciting Grounded Chain-of-Thought Reasoning in 3D Scenes

arXiv 2510.16714 ↗First posted 19 Oct 2025cs.CV cs.AIcited by 8

On the map: LLM reasoning → Spatial and embodied reasoning

Abstract (arXiv)

Existing research on 3D Large Language Models (LLMs) still struggles to achieve grounded question-answering, primarily due to the under-exploration of the mechanism of human-like scene-object grounded reasoning. This paper bridges the gap by presenting a novel framework. We first introduce a grounded Chain-of-Thought reasoning method in 3D scenes (SCENECOT), decoupling a complex reasoning task into simpler and manageable problems, and building corresponding visual clues based on multimodal expert modules. To enable such a method, we develop SCENECOT-185K, the first large-scale grounded CoT reasoning dataset, consisting of 185K high-quality instances. Extensive experiments across various complex 3D scene reasoning benchmarks demonstrate that our new framework achieves strong performance with high grounding-QA coherence. To the best of our knowledge, this is the first successful application of CoT reasoning to 3D scene understanding, enabling step-by-step human-like reasoning and showing potential for extension to broader 3D scene understanding scenarios.

Key facts. Cited by 8 papers in the map (data through 14 Oct 2026). Placed in Spatial and embodied reasoning, one of 17 directions in LLM reasoning. Builds on 39 of its 50 references that are in the map. First posted on arXiv 19 Oct 2025.

Builds on

39 of 50 references.

Closest ancestors · 7

Same direction, newest first

  1. 2025SpatialCoT: Advancing Spatial Reasoning through Coordinate Alignment and Chain-of-Thought for Embodied Task Planning17 Jan
  2. 2024Thinking in Space: How Multimodal Large Language Models See, Remember, and Recall Spaces18 Dec
  3. 2024Multi-modal Situated Reasoning in 3D Scenes4 Sep
  4. 2024Scene-LLM: Extending Language Model for 3D Visual Understanding and Reasoning18 Mar
  5. 2024SpatialVLM: Endowing Vision-Language Models with Spatial Reasoning Capabilities22 Jan
  6. 2023LL3DA: Visual Interactive Instruction Tuning for Omni-3D Understanding, Reasoning, and Planning30 Nov
  7. 2022SQA3D: Situated Question Answering in 3D Scenes14 Oct

Other references · 32

  1. 2023GPT-4 Technical Reportcited by 18,396
  2. 2022Chain-of-Thought Prompting Elicits Reasoning in Large Language Modelscited by 12,459
  3. 2024DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Modelscited by 9,379
  4. 2023Visual Instruction Tuningcited by 7,869
  5. 20233D Gaussian Splatting for Real-Time Radiance Field Renderingcited by 6,226
  6. 2022Flamingo: a Visual Language Model for Few-Shot Learningcited by 4,386
  7. 2023Kosmos-2: Grounding Multimodal Large Language Models to the Worldcited by 1,089
  8. 2023Shikra: Unleashing Multimodal LLM's Referential Dialogue Magiccited by 906
  9. 2023V*: Guided Visual Search as a Core Mechanism in Multimodal LLMscited by 482
  10. 2023An Embodied Generalist Agent in 3D Worldcited by 365
  11. 2023PointLLM: Empowering Large Language Models to Understand Point Cloudscited by 333
  12. 2023Uni3D: Exploring Unified 3D Representation at Scalecited by 237
20 more
  1. 2024LLaVA-3D: A Simple yet Effective Pathway to Empowering LMMs with 3D-awarenesscited by 183
  2. 2023Chat-Scene: Bridging 3D Scene and Large Language Models with Object Identifierscited by 175
  3. 2024Video-of-Thought: Step-by-Step Video Reasoning from Perception to Cognitioncited by 153
  4. 2024Video-3D LLM: Learning Position-Aware Video Representation for 3D Scene Understandingcited by 138
  5. 2024RoboSpatial: Teaching Spatial Understanding to 2D and 3D Vision-Language Models for Roboticscited by 136
  6. 2024SceneVerse: Scaling 3D Vision-Language Learning for Grounded Scene Understandingcited by 133
  7. 2024OMG-LLaVA: Bridging Image-level, Object-level, Pixel-level Reasoning and Understandingcited by 127
  8. 2024Grounded 3D-LLM with Referent Tokenscited by 115
  9. 2024ChatScene: Knowledge-Enabled Safety-Critical Scenario Generation for Autonomous Vehiclescited by 93
  10. 2019Holistic++ Scene Understanding: Single-view 3D Holistic Scene Parsing and Human Pose Estimation with Human-Object Interaction and Physical Commonsensecited by 71
  11. 2023GPT4Point: A Unified Framework for Point-Language Understanding and Generationcited by 71
  12. 2024Unifying 3D Vision-Language Understanding via Promptable Queriescited by 70
  13. 2022UniT3D: A Unified Transformer for 3D Dense Captioning and Visual Groundingcited by 69
  14. 2023ARNOLD: A Benchmark for Language-Grounded Task Learning With Continuous States in Realistic 3D Scenescited by 69
  15. 20243D-GRAND: A Million-Scale Dataset for 3D-LLMs with Better Grounding and Less Hallucinationcited by 36
  16. 2025SplatTalk: 3D VQA with Gaussian Splattingcited by 24
  17. 2025Unveiling the Mist over 3D Vision-Language Understanding: Object-centric Evaluation with Chain-of-Analysiscited by 23
  18. 2022Perceive, Ground, Reason, and Act: A Benchmark for General-purpose Visual Representationcited by 7
  19. 2024Unified Scene Representation and Reconstruction for 3D Large Language Modelscited by 5
  20. 2025LEO-VL: Efficient Scene Representation for Scalable 3D Vision-Language Learningcited by 5

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 1; 2026: 7.

8 papers.

  1. 2025D3D-VLP: Dynamic 3D Vision-Language-Planning Model for Embodied Grounding and Navigationcited by 5
  2. 2026Dense Reward for Multi-View 3D Reasoning with Global Maps and Local Viewscited by 3
  3. 2026Shape of Thought: Progressive Object Assembly via Visual Chain-of-Thoughtcited by 1
  4. 20263D-RFT: Reinforcement Fine-Tuning for Video-based 3D Scene Understandingcited by 1
  5. 2026SportsGrounder: Proposal-Aided Interleaved Grounding for Dense Sports Video Reasoningcited by 0
  6. 2026NeuSOGA3D: A Neuro-Symbolic Framework for Explainable 3D Geometric Reconstructioncited by 0
  7. 2026From Reasoning Failures to Composable Video Spatial Intelligencecited by 0
  8. 2026EgoAlign: Situated 3D Understanding without Situation Descriptionscited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2025Scene-R1: Video-Grounded Large Language Models for 3D Scene Reasoning without 3D Annotations
  2. 2024SceneGPT: A Language Model for 3D Scene Understanding
  3. 2026SceneBench: A Hierarchical Benchmark for Vision-Language Understanding of 3D Scenes
  4. 2024ScanReason: Empowering 3D Visual Grounding with Reasoning Capabilities
  5. 2025Hypo3D: Exploring Hypothetical Reasoning in 3D
  6. 2026PointCoT: A Multi-modal Benchmark for Explicit 3D Geometric Reasoning
  7. 2026Chat-Scene++: Exploiting Context-Rich Object Identification for 3D LLM
  8. 2025Grounded Chain-of-Thought for Multimodal Large Language Models