mlsanityData to 7 Oct 2026

Embodied and 3D vision-language models

2,288 papers8 this weekLow confidence in this grouping: the papers may belong to more than one topic, or the group may changeLow confidence in this grouping: the papers may belong to more than one topic, or the group may change · About

Keywords: reasoning, spatial, vlms, language, vision-language

Key facts. Embodied and 3D vision-language models: 2,288 papers on arXiv since 2021 (data through 7 Oct 2026). 256 of its 2,288 mapped papers (11%) cite “SpatialVLM: Endowing Vision-Language Models with Spatial Reasoning Capabilities” (2024, arXiv 2401.12168). 190 of its 2,288 mapped papers (8%) cite “Thinking in Space: How Multimodal Large Language Models See, Remember, and Recall Spaces” (2024, arXiv 2412.14171). 85 new papers 17 Sep–7 Oct.

What’s new

17 Sep – 7 Oct 2026

85 new papers, 17 Sep – 7 Oct; 8 this week.

  1. This week · 8
  2. VICO: Visual Environments Co-Evolving for Vision-Language Model Reasoning
  3. SpaTime: Streaming Vision-Language Models for Spatio-temporal Reasoning
  4. Selective Transfer of RL Updates for Visual Reasoning also fits: VLM benchmarks and visual reasoning
  5. AssemState: Manual and Physical-State-Guided Reasoning for Zero-shot Furniture Assembly
  6. TacZero: Training-Free Peg Insertion Using a General-Purpose Vision-Language Model with Tactile Feedback
  7. SpatialChain: A Benchmark for Auditing Spatial Reasoning Faithfulness in VLMs also fits: VLM benchmarks and visual reasoning
  8. Rotated, but How Far? Diagnosing and Improving Object-Rotation Reasoning in VLMs also fits: VLM benchmarks and visual reasoning
  9. StageVLN: Spatial and Trajectory Auxiliary Guidance for Efficient Vision-Language Navigation
  10. Earlier · 77
  11. Render to Reason: Novel-View Semantic Prediction Improves Spatial Understanding in VLMs
  12. RMMBench: A Comprehensive Benchmark for Robotic Mobile Manipulation
All 85
  1. IRSTD-Agent: Agentic Infrared Small Target Detection via Zoom-Guided Interaction Learning
  2. Recurrent Latent Visual Search for GUI Grounding also fits: VLM benchmarks and visual reasoning
  3. GeoBridge-VLA: Geometry-Aware Residual Adaptation for Vision-Language-Action Models
  4. BoT-Feedback: Grounding Multimodal Reasoning in Biomechanical Evidence for Explainable Human Action Feedback
  5. EgoExo-Next:Benchmarking Vision-Language Models on Visual-Option Next-State and Cross-View Reasoning also fits: Video understanding with MLLMs
  6. From Sight to Foresight: Predictive Spatial Reasoning in Vision-Language Models
  7. Foresight: planning future perception in streaming VLMs without retraining
  8. Seeing, Saying, but Not Using: From Reportable Spatial Facts to Usable States in Multimodal Large Language Models
  9. DuoMind: Enabling Distributed Multi-Robot Coordination with Semantic Communication
  10. GeoLatent: Geometry-Guided Latent Structuring with Routed Optimization for 3D Reasoning
  11. UniWAM: Unified World-Action Model
  12. Task-Adaptive Grounded 3D-Programmers Using 2D VLMs
  13. From Reasoning Failures to Composable Video Spatial Intelligence
  14. Fewer Tokens, Better Action: GPT-6 Astra Robot Agents with 14% Higher Success Rate but 65% Fewer Tokens
  15. Lang3DSeg: Annotation-Free Open-Vocabulary 3D Segmentation with Point Transformers
  16. Are Frontier VLM Agents Ready to Be Robot Generalists? An Empirical Study with the Embodied Agent Arena
  17. Toward Controlling Biology with Language:Offline Learning of Prompt-Conditioned Interventions for Cells, Organoids, and Biobots
  18. Token-World: World Modeling in Vision-Language Model Token Space for Robot Manipulation
  19. PhysVista: Benchmarking Physical Intelligence in VLMs via a Perception-Reasoning-Assessment Loop
  20. ChronoGraph: Functional 4D Scene Graphs with Vision-Language Models for Interaction Understanding and Grounded Planning
  21. KilometerVision: A New Frontier for Large-Scale Spatial Intelligence in VLMs
  22. Learning Reliable GUI Agents under Imperfect Priors
  23. Refusals That Bend: Measuring and Predicting Task Malleability in Embodied VLM Planners
  24. Soft Spatial Reasoning
  25. SpatialCORE: Confidence-Aware Grounded Spatial Reasoning in Large Vision--Language Models
  26. RoboHarn-Evo: Evolving Hierarchical Physical Knowledge for Self-Improving Robotic Manipulation
  27. Visual sensitivity is not claim retractability: persistence-aware credit assignment for multimodal reinforcement learning also fits: VLM benchmarks and visual reasoning
  28. GaugeVLM: Structuring Spatial Supervision with Measured Geometric Interventions
  29. Imagine3D-LLM: Teaching MLLMs to Imagine 3D Scenes Before Answering also fits: VLM benchmarks and visual reasoning
  30. MotorMind: Scaffolding General Vision Language Models for Zero-Shot Robot Manipulation
  31. Exemplar2VQA: A Scalable Exemplar-Driven Visual Question Answering Generation Framework via Multi-Agent Coding
  32. Spatial-OPSD: Self-Improving Spatial Reasoning via Label-Free Self-Distillation
  33. SphMind: Towards Robust, Training-Free VLM-based Spatial Reasoning with a 360 Camera
  34. DualManip: Agentic Dynamic Manipulation via Dual-Path Semantic Reasoning and Geometric Adaptation
  35. Verifiable Visual Rewards Transfer from Synthetic Scenes to Natural Prompts also fits: VLM benchmarks and visual reasoning
  36. NavJev: Efficient Vision-Language Navigation via Action-Centric Visual Compression and Discriminative Action-Semantic Memory
  37. PanoVLN: Towards Effective Panoramic Vision-and-Language Navigation
  38. Reinforcement Learning from Intermediate Renders for Image-to-Code Generation also fits: Image editing and quality assessment
  39. RoboICL: Embodied In-Context Learning with GPT-6 Astra
  40. Geometric Encoding for Spatial Reasoning in Vision-Language Models
  41. SpatialSkill: Self-Evolving Skills for Cross-View Spatial Reasoning
  42. PhysFieldBench: Can Multimodal Models Understand Physical Fields? also fits: MLLM training and visual instruction tuning
  43. Quantile Head for Vision-Language-Action Models
  44. Does Adversarial Training Improve Generalization in Multi-View VLAs? Revealing and Mitigating View Collapse also fits: MLLM jailbreaks and adversarial robustness
  45. SpatialSpeak: QA-Native Reconstruction with Local and Global Context for Spatial Chain-of-Thought Reasoning
  46. SceneScaffold: Active Scene-State Construction for Unified 3D Scene Understanding also fits: VLM benchmarks and visual reasoning
  47. Toward Comprehensive 3D Grounding: Orientation Grounding through Vision-Language Models
  48. Vision-Language Agents for Active Perception in Optics Laboratories
  49. USAI-Quant: A Quantitative Reasoning Benchmark for Vision-Language Models in Built Environments also fits: VLM benchmarks and visual reasoning
  50. CT-OPD: Counterfactual Trace On-Policy Distillation for Diffusion Vision-Language Models
  51. Gradient-Guided Decoupled Adaptation for Geospatial Vision-Language Models also fits: Prompt learning and adaptation of CLIP
  52. CityToolVQA: Tool-Augmented Visual Question Answering for 3D Spatial Cognition in Urban Low-Altitude Environments also fits: VLM benchmarks and visual reasoning
  53. DepthEvidence: Unifying Metric Depth Prediction and Geometric Reasoning in Multimodal Language Models also fits: VLM benchmarks and visual reasoning
  54. World Action Agent: Harnessing VLMs for Robot Manipulation via World Action Rehearsal
  55. Robo-Harness K1: Harnessing Robot-Use Agents via Perception Augmentation
  56. Decoupled Early Exits for Task-Dependent Compute Allocation in Flow-Matching VLAs
  57. Seeing Is Not Measuring: Tool-Augmented Metric Spatial Reasoning for Vision-Language Models
  58. EgoTSR++: Egocentric Spatiotemporal Reasoning for Task Progress Understanding
  59. The Earth in One Gaze: Training-Free Active Focus for UHR Remote Sensing Understanding also fits: VLM benchmarks and visual reasoning
  60. Small yet Assistive: Spatially-Aware Post-Training for Low Vision also fits: Prompt learning and adaptation of CLIP
  61. Pistis Technical Report also fits: MLLM training and visual instruction tuning
  62. A Unified Framework and Dataset for Oriented Object Visual Grounding in Remote Sensing also fits: VLM benchmarks and visual reasoning
  63. Know Your Body: A Harness for Direct and Self-Improving Robot Control with VLMs
  64. Metric-Bench: Exploring In-context Spatial Metric Reasoning in VLMs for Indoor Scenes
  65. INTCORT: Training-Free Spatial Reasoning Enhancement for Vision-Language Models via Input Transformations and Confidence Routing
  66. What do VLM-Based Vision-Language Navigation Models Rely on: Interpreting and Steering Policy Behavior
  67. VLM-in-Sandbox: Visual Workspaces for Agentic Visual Reasoning also fits: VLM benchmarks and visual reasoning
  68. Taming CoT Obfuscation in VLMs: From Mechanistic Evidence to Activation Enforcement also fits: MLLM jailbreaks and adversarial robustness
  69. Spatial-Interactor: Learning Spatial Reasoning through Interaction with the Observable Physical World
  70. "Dear LLaVA, Please Drive": A Depth-Aware Vision-Language Agent for Closed-Loop Robotic Control also fits: Driving VLMs and open-vocabulary detection
  71. MintAct: A Unified Visual Agent for Digital Environments
  72. Scaling Vision-Language Reward Learning for Robot Manipulation in Parallel Simulation also fits: Prompt learning and adaptation of CLIP
  73. DRT: Dense Reasoning Trace for Efficient and Grounded Multimodal Reasoning also fits: VLM benchmarks and visual reasoning
  74. FOCAL-VLA: Subtask-Guided Geometry Distillation and Implicit World Modeling for Vision-Language-Action Models
  75. Catch Me If You Can: Real-Time Feedback Denoising for Responsive VLAs

Foundations

Most cited by papers here

  1. SpatialVLM: Endowing Vision-Language Models with Spatial Reasoning Capabilities2024cited by 256 here
  2. Thinking in Space: How Multimodal Large Language Models See, Remember, and Recall Spaces2024cited by 190 here
  3. SpatialRGPT: Grounded Spatial Reasoning in Vision Language Models2024cited by 161 here
  4. RT-2: Vision-Language-Action Models Transfer Web Knowledge to Robotic Control2023cited by 157 here
  5. PaLM-E: An Embodied Multimodal Language Model2023cited by 148 here

Lineage

The papers this direction builds on, oldest first.

Each paper on the line cites the one before.

  1. 2022Flamingo: a Visual Language Model for Few-Shot Learningcited by 199 here
  2. 2023PaLM-E: An Embodied Multimodal Language Modelcited by 148 here
  3. 2023InstructBLIP: Towards General-purpose Vision-Language Models with Instruction Tuningcited by 132 here
  4. 2023Qwen-VL: A Versatile Vision-Language Model for Understanding, Localization, Text Reading, and Beyondcited by 142 here
  5. 2024Measuring Multimodal Mathematical Reasoning with MATH-Vision Datasetcited by 112 here
  6. 2024Qwen2-VL: Enhancing Vision-Language Model's Perception of the World at Any Resolutioncited by 226 here
  7. 2024Expanding Performance Boundaries of Open-Source Multimodal Models with Model, Data, and Test-Time Scalingcited by 132 here
17 more it builds on, not in this chain
  1. 2023MMBench: Is Your Multi-modal Model an All-around Player?cited by 128 here
  2. 20233D-LLM: Injecting the 3D World into Large Language Modelscited by 117 here
  3. 2023RT-2: Vision-Language-Action Models Transfer Web Knowledge to Robotic Controlcited by 157 here
  4. 2023MathVista: Evaluating Mathematical Reasoning of Foundation Models in Visual Contextscited by 173 here
  5. 2023Improved Baselines with Visual Instruction Tuningcited by 214 here
  6. 2023GeoChat: Grounded Large Vision-Language Model for Remote Sensingcited by 102 here
  7. 2023InternVL: Scaling up Vision Foundation Models and Aligning for Generic Visual-Linguistic Taskscited by 189 here
  8. 2024SpatialVLM: Endowing Vision-Language Models with Spatial Reasoning Capabilitiescited by 256 here
  9. 2024MathVerse: Does Your Multi-modal LLM Truly See the Diagrams in Visual Math Problems?cited by 121 here
  10. 2024BLINK: Multimodal Large Language Models Can See but Not Perceivecited by 108 here
  11. 2024SpatialRGPT: Grounded Spatial Reasoning in Vision Language Modelscited by 161 here
  12. 2024Cambrian-1: A Fully Open, Vision-Centric Exploration of Multimodal LLMscited by 111 here
  13. 2024LLaVA-OneVision: Easy Visual Task Transfercited by 227 here
  14. 2024Thinking in Space: How Multimodal Large Language Models See, Remember, and Recall Spacescited by 190 here
  15. 2025Vision-R1: Incentivizing Reasoning Capability in Multimodal Large Language Modelscited by 119 here
  16. 2025VLM-R1: A Stable and Generalizable R1-style Large Vision-Language Modelcited by 105 here
  17. 2025Qwen3-VL Technical Reportcited by 170 here

The 24 most cited of 2,579.

Browse

30 of 2,288, ranked 6–35

  1. 6Vision-R1: Incentivizing Reasoning Capability in Multimodal Large Language Models2025119
  2. 73D-LLM: Injecting the 3D World into Large Language Models2023117
  3. 8VLM-R1: A Stable and Generalizable R1-style Large Vision-Language Model2025105
  4. 9GeoChat: Grounded Large Vision-Language Model for Remote Sensing2023102
  5. 10SpatialBot: Precise Spatial Understanding with Vision Language Models202484
  6. 11Spatial-MLLM: Boosting MLLM Capabilities in Visual-based Spatial Intelligence202578
  7. 12What's "up" with vision-language models? Investigating their struggle with spatial reasoning202376
  8. 13VL-Rethinker: Incentivizing Self-Reflection of Vision-Language Models with Reinforcement Learning202575
  9. 14VLM-3R: Vision-Language Models Augmented with Instruction-Aligned 3D Reconstruction202574
  10. 15RSGPT: A Remote Sensing Vision Language Model and Benchmark202373
All 30
  1. 16R1-VL: Learning to Reason with Multimodal Large Language Models via Step-wise Group Relative Policy Optimization202571
  2. 17EarthGPT: A Universal Multi-modal Large Language Model for Multi-sensor Image Comprehension in Remote Sensing Domain202463
  3. 18LHRS-Bot: Empowering Remote Sensing with VGI-Enhanced Large Multimodal Language Model202461
  4. 19LLaVA-3D: A Simple yet Effective Pathway to Empowering LMMs with 3D-awareness202460
  5. 20RoboSpatial: Teaching Spatial Understanding to 2D and 3D Vision-Language Models for Robotics202460
  6. 21EmbSpatial-Bench: Benchmarking Spatial Understanding for Embodied Tasks with Large Vision-Language Models202459
  7. 22Video-3D LLM: Learning Position-Aware Video Representation for 3D Scene Understanding202458
  8. 23LL3DA: Visual Interactive Instruction Tuning for Omni-3D Understanding, Reasoning, and Planning202358
  9. 24SpaceR: Reinforcing MLLMs in Video Spatial Reasoning202557
  10. 25DeepEyes: Incentivizing "Thinking with Images" via Reinforcement Learning202557
  11. 26From Flatland to Space: Teaching Vision-Language Models to Perceive and Reason in 3D202555
  12. 27MMSI-Bench: A Benchmark for Multi-Image Spatial Intelligence202553
  13. 28OmniSpatial: Towards Comprehensive Spatial Reasoning Benchmark for Vision Language Models202553
  14. 29SFT or RL? An Early Investigation into Training R1-Like Reasoning Large Vision-Language Models202552
  15. 30VRSBench: A Versatile Vision-Language Benchmark Dataset for Remote Sensing Image Understanding202451
  16. 31RemoteCLIP: A Vision Language Foundation Model for Remote Sensing202351
  17. 32GLM-4.5V and GLM-4.1V-Thinking: Towards Versatile Multimodal Reasoning with Scalable Reinforcement Learning202550
  18. 33Gemini Robotics: Bringing AI into the Physical World202547
  19. 34Reinforcing Spatial Reasoning in Vision-Language Models with Interwoven Thinking and Visual Drawing202547
  20. 35SAT: Dynamic Spatial Aptitude Training for Multimodal Language Models202447