mlsanityData to 14 Oct 2026

ThinkMorph: Emergent Properties in Multimodal Interleaved Chain-of-Thought Reasoning

arXiv 2510.27492 ↗First posted 30 Oct 2025cs.CVcited by 56

On the map: LLM reasoning → Multimodal reasoning

Abstract (arXiv)

Multimodal reasoning requires iterative coordination between language and vision, yet it remains unclear what constitutes a meaningful interleaved chain of thought. We posit that text and image thoughts should function as complementary rather than isomorphic modalities that mutually advance reasoning. Guided by this principle, we build ThinkMorph, a unified model fine-tuned on approximately 24K high-quality interleaved reasoning traces spanning tasks with varying visual engagement. ThinkMorph learns to generate progressive text-image reasoning steps that concretely manipulate visual content while maintaining coherent verbal logic. It delivers large gains on vision-centric benchmarks (averaging 34.7 percent over the base model) and generalizes to out-of-domain tasks, matching or surpassing larger and proprietary VLMs. Beyond performance, ThinkMorph exhibits emergent multimodal intelligence, including unseen visual manipulation skills, adaptive switching between reasoning modes, and better test-time scaling through diversified multimodal thoughts. These findings suggest promising directions for characterizing the emergent capabilities of unified models for multimodal reasoning.

Key facts. Cited by 56 papers in the map (data through 14 Oct 2026). Placed in Multimodal reasoning, one of 17 directions in LLM reasoning. Builds on 34 of its 49 references that are in the map. First posted on arXiv 30 Oct 2025.

Builds on

34 of 49 references.

Closest ancestors · 16

Same direction, newest first

  1. 2025Reinforced Visual Perception with Tools1 Sep
  2. 2025InternVL3.5: Advancing Open-Source Multimodal Models in Versatility, Reasoning, and Efficiency25 Aug
  3. 2025Uni-cot: Towards Unified Chain-of-Thought Reasoning Across Text and Vision7 Aug
  4. 2025Zebra-CoT: A Dataset for Interleaved Vision Language Reasoning22 Jul
  5. 2025Machine Mental Imagery: Empower Multimodal Reasoning with Latent Visual Tokens20 Jun
  6. 2025Thinking with Generated Images28 May
  7. 2025Pixel Reasoner: Incentivizing Pixel-Space Reasoning with Curiosity-Driven Reinforcement Learning21 May
  8. 2025DeepEyes: Incentivizing "Thinking with Images" via Reinforcement Learning20 May
  9. 2025Mind with Eyes: from Language Reasoning to Multimodal Reasoning23 Mar
  10. 2025MME-CoT: Benchmarking Chain-of-Thought in Large Multimodal Models for Reasoning Quality, Robustness, and Efficiency13 Feb
  11. 2025Imagine while Reasoning in Space: Multimodal Visualization-of-Thought13 Jan
  12. 2025ReFocus: Visual Editing as a Chain of Thought for Structured Image Understanding9 Jan
4 more
  1. 2025Can MLLMs Reason in Multimodality? EMMA: An Enhanced MultiModal ReAsoning Benchmark9 Jan
  2. 2024Image-of-Thought Prompting for Visual Reasoning Refinement in Multimodal Large Language Models22 May
  3. 2024Visual CoT: Advancing Multi-Modal Language Models with a Comprehensive Dataset and Benchmark for Chain-of-Thought Reasoning25 Mar
  4. 2022ChartQA: A Benchmark for Question Answering about Charts with Visual and Logical Reasoning19 Mar

Other references · 18

  1. 2025Qwen2.5-VL Technical Reportcited by 5,587
  2. 2016OpenAI Gymcited by 1,736
  3. 2024Chameleon: Mixed-Modal Early-Fusion Foundation Modelscited by 927
  4. 2025Emerging Properties in Unified Multimodal Pretrainingcited by 897
  5. 2024BLINK: Multimodal Large Language Models Can See but Not Perceivecited by 550
  6. 2025Transfer between Modalities with MetaQueriescited by 236
  7. 2024MetaMorph: Multimodal Understanding and Generation via Instruction Tuningcited by 206
  8. 2024ANOLE: An Open, Autoregressive, Native Large Multimodal Models for Interleaved Image-Text Generationcited by 114
  9. 2025MindCube: Spatial Mental Modeling from Limited Viewscited by 95
  10. 2024VSP: Assessing the dual challenges of perception and reasoning in spatial planning tasks for VLMscited by 47
  11. 2025Interleaving Reasoning for Better Text-to-Image Generationcited by 41
  12. 2024IsoBench: Benchmarking Multimodal Foundation Models on Isomorphic Representationscited by 38
6 more
  1. 2025Jigsaw-R1: A Study of Rule-based Visual Reinforcement Learning with Jigsaw Puzzlescited by 36
  2. 2025Holistic Evaluation of Multimodal LLMs on Spatial Intelligencecited by 29
  3. 2025Introducing Visual Perception Token into Multimodal Large Language Modelcited by 26
  4. 2025Unfolding Spatial Cognition: Evaluating Multimodal Models on Visual Simulationscited by 21
  5. 2025Autoregressive Semantic Visual Reconstruction Helps VLMs Understand Bettercited by 17
  6. 2025Unified Multimodal Models as Auto-Encoderscited by 13

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 6; 2026: 50.

56 papers.

Sort:

By area:

  1. 2025VisMem: Latent Vision Memory Unlocks Potential of Vision-Language Modelscited by 37
  2. 2026SenseNova-U1: Unifying Multimodal Understanding and Generation with NEO-unify Architecturecited by 32
  3. 2025TiViBench: Benchmarking Think-in-Video Reasoning for Video Generative Modelscited by 19
  4. 2025Mull-Tokens: Modality-Agnostic Latent Thinkingcited by 19
  5. 2025DiffThinker: Towards Generative Multimodal Reasoning with Diffusion Modelscited by 19
  6. 2026Visual Generation Unlocks Human-Like Reasoning through Multimodal World Modelscited by 19
  7. 2026Visual-OPSD: Cross-Modal On-Policy Self-Distillation for Efficient Unified Multimodal Reasoningcited by 16
  8. 2026LaViT: Aligning Latent Visual Thoughts for Multi-modal Reasoningcited by 15
  9. 2026Thinking in Frames: How Visual Context and Test-Time Scaling Empower Video Reasoningcited by 10
  10. 2025COOPER: A Unified Model for Cooperative Perception and Reasoning in Spatial Intelligencecited by 9
  11. 2026Show, Don't Tell: Morphing Latent Reasoning into Image Generationcited by 8
  12. 2026UniT: Unified Multimodal Chain-of-Thought Test-time Scalingcited by 8
38 more
  1. 2026LatentUM: Unleashing the Potential of Interleaved Cross-Modal Reasoning via a Latent-Space Unified Modelcited by 8
  2. 2026SenseNova-U1.5: Towards Native Unified Visual Intelligencecited by 7
  3. 2025IRG-MotionLLM: Interleaving Motion Generation, Assessment and Refinement for Text-to-Motion Generationcited by 6
  4. 2026VBVR-Pro: A Scalable and Verifiable Suite for Native Visual Reasoningcited by 6
  5. 2026Bridging Interleaved Multi-Modal Reasoning as a Unified Decision Processcited by 5
  6. 2026SwimBird: Eliciting Switchable Reasoning Mode in Hybrid Autoregressive MLLMscited by 4
  7. 2026Multimodal Latent Reasoning via Predictive Embeddingscited by 4
  8. 2026Unveiling Fine-Grained Visual Traces: Evaluating Multimodal Interleaved Reasoning Chains in Multimodal STEM Taskscited by 4
  9. 2026CollabVR: Collaborative Video Reasoning with Vision-Language and Video Generation Modelscited by 4
  10. 2026DuoGen: Towards General Purpose Interleaved Multimodal Generationcited by 3
  11. 2026TwiFF (Think With Future Frames): A Large-Scale Dataset for Dynamic Visual Reasoningcited by 3
  12. 2026StruVis: Enhancing Reasoning-based Text-to-Image Generation via Thinking with Structured Visioncited by 3
  13. 2026ProLaViT: Learning Progressive Latent Visual Thoughts in Structured Latent Spacecited by 3
  14. 2026How RL Unlocks the Aha Moment in Geometric Interleaved Reasoningcited by 2
  15. 2026ETCHR: Editing To Clarify and Harness Reasoningcited by 2
  16. 2026OmniVerifier-M1: Multimodal Meta-Verifier with Explicit Structured Recalibrationcited by 2
  17. 2026World Models Meet Language Models: On the Complementarity of Concrete and Abstract Reasoningcited by 2
  18. 2026RL Forgets! Towards Continual Policy Optimizationcited by 2
  19. 2026Scaffolding Minds: Optimizing Latent Visual Target Representations for Multimodal Reasoningcited by 2
  20. 2026Shape of Thought: Progressive Object Assembly via Visual Chain-of-Thoughtcited by 1
  21. 2026Benchmarking and Evolving Reason-Reflect-Rectify for Reflective Visual Generationcited by 1
  22. 2026Entropy Is Not Enough: Unlocking Effective Reinforcement Learning for Visual Reasoning via Vision-Anchored Token Selectioncited by 1
  23. 2026Imaginative Perception Tokens Enhance Spatial Reasoning in Multimodal Language Modelscited by 1
  24. 2026DeltaV: Thinking with Visual State Updates in Unified Large Multimodal Modelscited by 1
  25. 2026Think, Plan, Paint: Layout-Aware Reasoning for Controllable Image Generation in Unified Modelscited by 1
  26. 2026SpatialImaginer: Towards Adaptive Visual Imagination for Spatial Reasoningcited by 0
  27. 2026SketchVLM: Vision language models can annotate images to explain thoughts and guide userscited by 0
  28. 2026Draw2Think: Harnessing Geometry Reasoning through Constraint Engine Interactioncited by 0
  29. 2026GeoWeaver: Grounding Visual Tokens with Geometric Evidence before Scene Reasoningcited by 0
  30. 2026InterSketch: An Interleaved Reasoning Model with Self-correcting Visual Sketch and Stepwise Rewardcited by 0
  31. 2026UniCanvas: A Diffusion-base Unified Model for Text-in-Image Joint Generationcited by 0
  32. 2026Bridging Modal Isolation in Interleaved Thinking: Supervising Modality Transitions via Stepwise Reinforcementcited by 0
  33. 2026Efficient Reinforcement for Visual-Textual Thinking with Discrete Diffusion Modelcited by 0
  34. 2026APIVOT: Adaptive Planning with Interleaved Vision-Language Thoughtscited by 0
  35. 2026UniVR: Thinking in Visual Space for Unified Visual Reasoningcited by 0
  36. 2026STBridge: Shared-Target Alignment for Bridging Understanding and Generation in UMMscited by 0
  37. 2026Aphanta: Diagnosing Task-Aligned Image-Edited Intermediates for Multimodal Reasoningcited by 0
  38. 2026MindTopo: Can Foundation Models Reason in Topological Space?cited by 0

50 most cited of 56 All 56 (JSON)

Similar, not cited

Close in wording, no citation either way

  1. 2025Reasoning Within the Mind: Dynamic Multimodal Interleaving in Latent Space
  2. 2025Visual Thoughts: A Unified Perspective of Understanding Multimodal Chain-of-Thought
  3. 2026Adaptive Look, Flexible Thought: Learning How and How Long to Think in Multimodal Reasoning
  4. 2025Thinking with Images for Multimodal Reasoning: Foundations, Methods, and Future Frontiers
  5. 2025Thinking with Visual Abstract: Enhancing Multimodal Reasoning via Visual Abstraction
  6. 2026Distilling Visual Reasoning into Text Space
  7. 2026UniVLR: Unifying Text and Vision in Visual Latent Reasoning for Multimodal LLMs
  8. 2026Look Light, Think Heavy: What Multimodal Chain-of-Thought Reasoning Can and Cannot Do