mlsanityData to 14 Oct 2026

SpaceR: Reinforcing MLLMs in Video Spatial Reasoning

arXiv 2504.01805 ↗First posted 2 Apr 2025cs.CVcited by 159

On the map: LLM reasoning → Spatial and embodied reasoning

also in 1 more
Abstract (arXiv)

Video spatial reasoning, which involves inferring the underlying spatial structure from observed video frames, poses a significant challenge for existing Multimodal Large Language Models (MLLMs). This limitation stems primarily from 1) the absence of high-quality datasets for this task, and 2) the lack of effective training strategies to develop spatial reasoning capabilities. Motivated by the success of Reinforcement Learning with Verifiable Reward (RLVR) in unlocking LLM reasoning abilities, this work aims to improve MLLMs in video spatial reasoning through the RLVR paradigm. To this end, we introduce the SpaceR framework. First, we present SpaceR-151k, a dataset with 91k questions spanning diverse spatial reasoning scenarios with verifiable answers, and 60k samples for maintaining general multimodal understanding. Second, we propose Spatially-Guided RLVR (SG-RLVR), a novel reinforcement learning approach that extends Group Relative Policy Optimization (GRPO) with a novel map imagination mechanism, which encourages the model to infer spatial layouts in the thinking process, thereby facilitating more effective spatial reasoning. Extensive experiments demonstrate that SpaceR achieves state-of-the-art performance on spatial reasoning benchmarks (e.g., VSI-Bench, STI-Bench, and SPAR-Bench), while maintaining competitive results on video understanding benchmarks (e.g., Video-MME, TempCompass, and LongVideoBench). Remarkably, SpaceR surpasses the advanced GPT-4o by 11.6% accuracy on VSI-Bench and is on par with the leading proprietary model Gemini-2.0-Flash, highlighting the effectiveness of our SpaceR-151k dataset and SG-RLVR in reinforcing spatial reasoning ability of MLLMs. Code, model, and dataset are available at https://github.com/OuyangKun10/SpaceR.

Key facts. Cited by 159 papers in the map (data through 14 Oct 2026). Placed in Spatial and embodied reasoning, one of 17 directions in LLM reasoning. Builds on 34 of its 38 references that are in the map. First posted on arXiv 2 Apr 2025.

Builds on

34 of 38 references.

Closest ancestors · 1

Same direction, newest first

  1. 2024Thinking in Space: How Multimodal Large Language Models See, Remember, and Recall Spaces18 Dec

Other references · 33

  1. 2024DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Modelscited by 9,379
  2. 2014ImageNet Large Scale Visual Recognition Challengecited by 8,498
  3. 2025Qwen2.5-VL Technical Reportcited by 5,587
  4. 2024Qwen2.5 Technical Reportcited by 4,942
  5. 2017ScanNet: Richly-annotated 3D Reconstructions of Indoor Scenescited by 3,117
  6. 2023Sigmoid Loss for Language Image Pre-Trainingcited by 3,109
  7. 2024LLaVA-OneVision: Easy Visual Task Transfercited by 2,719
  8. 2023InternVL: Scaling up Vision Foundation Models and Aligning for Generic Visual-Linguistic Taskscited by 2,565
  9. 2015VQA: Visual Question Answeringcited by 2,506
  10. 2023PaLM-E: An Embodied Multimodal Language Modelcited by 2,071
  11. 2024Expanding Performance Boundaries of Open-Source Multimodal Models with Model, Data, and Test-Time Scalingcited by 1,599
  12. 2024Video-MME: The First-Ever Comprehensive Evaluation Benchmark of Multi-modal LLMs in Video Analysiscited by 1,413
21 more
  1. 2015Flickr30k Entities: Collecting Region-to-Phrase Correspondences for Richer Image-to-Sentence Modelscited by 1,362
  2. 2025Kimi k1.5: Scaling Reinforcement Learning with LLMscited by 1,210
  3. 2023MVBench: A Comprehensive Multi-modal Video Understanding Benchmarkcited by 1,109
  4. 2024MiniCPM-V: A GPT-4V Level MLLM on Your Phonecited by 1,079
  5. 2016Modeling Context in Referring Expressionscited by 1,063
  6. 2024LongVideoBench: A Benchmark for Long-context Interleaved Video-Language Understandingcited by 665
  7. 2017TALL: Temporal Activity Localization via Language Querycited by 584
  8. 2025VideoLLaMA 3: Frontier Multimodal Foundation Models for Image and Video Understandingcited by 518
  9. 2025Visual-RFT: Visual Reinforcement Fine-Tuningcited by 513
  10. 2025Video-R1: Reinforcing Video Reasoning in MLLMscited by 484
  11. 2025Muon is Scalable for LLM Trainingcited by 428
  12. 2024TempCompass: Do Video LLMs Really Understand Videos?cited by 363
  13. 2023Perception Test: A Diagnostic Benchmark for Multimodal Video Modelscited by 314
  14. 2025LMM-R1: Empowering 3B LMMs with Strong Reasoning Abilities Through Two-Stage Rule-Based RLcited by 248
  15. 2024World Model on Million-Length Video And Language With Blockwise RingAttentioncited by 198
  16. 2025From Flatland to Space: Teaching Vision-Language Models to Perceive and Reason in 3Dcited by 138
  17. 2014Translating Videos to Natural Language Using Deep Recurrent Neural Networkscited by 95
  18. 2025STI-Bench: Are MLLMs Ready for Precise Spatial-Temporal World Understanding?cited by 79
  19. 2025Virgo: A Preliminary Exploration on Reproducing o1-like MLLMcited by 68
  20. 2022Embracing Consistency: A One-Stage Approach for Spatio-Temporal Video Groundingcited by 42
  21. 2023Diversifying Spatial-Temporal Perception for Video Domain Generalizationcited by 19

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 43; 2026: 116.

159 papers.

Sort:

By area:

  1. 2025Spatial-MLLM: Boosting MLLM Capabilities in Visual-based Spatial Intelligencecited by 217
  2. 2025The Landscape of Agentic Reinforcement Learning for LLMs: A Surveycited by 205
  3. 2025A Survey of Reinforcement Learning for Large Reasoning Modelscited by 180
  4. 2025Cambrian-S: Towards Spatial Supersensing in Videocited by 134
  5. 2025Perception, Reason, Think, and Plan: A Survey on Large Multimodal Reasoning Modelscited by 90
  6. 2025SpatialLadder: Progressive Training for Spatial Reasoning in Vision-Language Modelscited by 90
  7. 2025Visual Spatial Tuningcited by 89
  8. 2025Scaling Spatial Intelligence with Multimodal Foundation Modelscited by 82
  9. 2025Think with 3D: Geometric Imagination Grounded Spatial Reasoning from Limited Viewscited by 65
  10. 2025Open-o3-Video: Grounded Video Reasoning with Explicit Spatio-Temporal Evidencecited by 49
  11. 2025How Far are VLMs from Visual Spatial Intelligence? A Benchmark-Driven Perspectivecited by 43
  12. 2025Spatial-SSRL: Enhancing Spatial Understanding via Self-Supervised Reinforcement Learningcited by 41
38 more
  1. 2026VLM4VLA: Revisiting Vision-Language-Models in Vision-Language-Action Modelscited by 39
  2. 2026LLaVA-OneVision-2: Towards Next-Generation Perceptual Intelligencecited by 39
  3. 2025Reinforcement Learning Meets Large Language Models: A Survey of Advancements and Applications Across the LLM Lifecyclecited by 35
  4. 2025Multimodal Spatial Reasoning in the Large Model Era: A Survey and Benchmarkscited by 33
  5. 2025Why Do MLLMs Struggle with Spatial Understanding? A Systematic Analysis from Data to Architecturecited by 30
  6. 2025Geometrically-Constrained Agent for Spatial Reasoningcited by 26
  7. 2025SpaceMind: Camera-Guided Modality Fusion for Spatial Reasoning in Vision-Language Modelscited by 26
  8. 2025SpatialScore: Towards Comprehensive Evaluation for Spatial Intelligencecited by 24
  9. 2025Vlaser: Vision-Language-Action Model with Synergistic Embodied Reasoningcited by 23
  10. 2025Uni-MoE-2.0-Omni: Scaling Language-Centric Omnimodal Large Model with Advanced MoE, Training and Datacited by 21
  11. 2025SpaceVista: All-Scale Visual Spatial Reasoning from mm to kmcited by 19
  12. 2025Reasoning in Space via Grounding in the Worldcited by 19
  13. 2025Conan: Progressive Learning to Reason Like a Detective over Multi-Scale Visual Evidencecited by 18
  14. 2026PyVision-RL: Forging Open Agentic Vision Models via RLcited by 16
  15. 2026ReVSI: Rebuilding Visual Spatial Intelligence Evaluation for Accurate Assessment of VLM 3D Reasoningcited by 15
  16. 2025Reinforcement Fine-Tuning Powers Reasoning Capability of Multimodal Large Language Modelscited by 14
  17. 2025SiLVR: A Simple Language-based Video Reasoning Frameworkcited by 14
  18. 2025SpinBench: Perspective and Rotation as a Lens on Spatial Reasoning in VLMscited by 14
  19. 2026Thinking in Dynamics: How Multimodal Large Language Models Perceive, Track, and Reason Dynamics in Physical 4D Worldcited by 12
  20. 2026JoyAI-Image: Awaking Spatial Intelligence in Unified Multimodal Understanding and Generationcited by 12
  21. 2025SpatialDreamer: Incentivizing Spatial Reasoning via Active Mental Imagerycited by 11
  22. 2026On Robustness and Chain-of-Thought Consistency of RL-Finetuned VLMscited by 11
  23. 2026Spatial-TTT: Streaming Visual-based Spatial Intelligence with Test-Time Trainingcited by 11
  24. 2026S-Agent: Spatial Tool-Use Elicits Reasoning for Spatial Intelligencecited by 11
  25. 2025Is your VLM Sky-Ready? A Comprehensive Spatial Intelligence Benchmark for UAV Navigationcited by 10
  26. 2026MLLM-4D: Towards Visual-based Spatial-Temporal Intelligencecited by 10
  27. 2026VISD: Enhancing Video Reasoning via Structured Self-Distillationcited by 10
  28. 2026VideoChat3: Fully Open Video MLLM for Efficient and Generalist Video Understandingcited by 10
  29. 2025Enhancing Spatial Reasoning in Vision-Language Models via Chain-of-Thought Prompting and Reinforcement Learningcited by 9
  30. 2025Spatial-DISE: A Unified Benchmark for Evaluating Spatial Reasoning in Vision-Language Modelscited by 9
  31. 2025COOPER: A Unified Model for Cooperative Perception and Reasoning in Spatial Intelligencecited by 9
  32. 2026Generate, Filter, Control, Replay: A Comprehensive Survey of Rollout Strategies for LLM Reinforcement Learningcited by 9
  33. 2025Video-LMM Post-Training: A Deep Dive into Video Reasoning with Large Multimodal Modelscited by 8
  34. 2025TempR1: Improving Temporal Understanding of MLLMs via Temporal-Aware Multi-Task Reinforcement Learningcited by 8
  35. 2026Can LLMs See Without Pixels? Benchmarking Spatial Intelligence from Textual Descriptionscited by 7
  36. 2026Thinking with Spatial Code for Physical-World Video Reasoningcited by 7
  37. 2026Generation Models Know Space: Unleashing Implicit 3D Priors for Scene Understandingcited by 7
  38. 2025Video-STR: Reinforcing MLLMs in Video Spatio-Temporal Reasoning with Relation Graphcited by 6

50 most cited of 159 All 159 (JSON)

Similar, not cited

Close in wording, no citation either way

  1. 2025SVQA-R1: Reinforcing Spatial Reasoning in MLLMs via View-Consistent Reward Optimization
  2. 2025M2-Reasoning: Empowering MLLMs with Unified General and Spatial Reasoning
  3. 2026Video-MSR: Benchmarking Multi-hop Spatial Reasoning Capabilities of MLLMs
  4. 2025ST-Think: How Multimodal Large Language Models Reason About 4D Worlds from Ego-Centric Videos
  5. 2025SpatialThinker: Reinforcing Scene Graph-Grounded Spatial Reasoning via Dense Rewards
  6. 2026Faithful GRPO: Improving Visual Spatial Reasoning in Multimodal Language Models via Constrained Policy Optimization
  7. 2026ReTAC: Residual Test-Time Agentic Computation for Video Spatial Reasoning
  8. 2025Scaling and Beyond: Advancing Spatial Reasoning in MLLMs Requires New Recipes