[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"detail-sidebar-cat-0-en-105":3,"doc-seo-139166-105":59,"doc-detail-139166-en":130},{"code":4,"msg":5,"data":6},0,"success",[7,13,18,23,28,33,38,43,48,51,55],{"id":8,"doc_module":4,"doc_module_name":9,"category_name":10,"show_sort_weight":11,"slug":12},1,"Document","Story & Novel",90,"story-novel",{"id":14,"doc_module":4,"doc_module_name":9,"category_name":15,"show_sort_weight":16,"slug":17},2,"Literature",80,"literature",{"id":19,"doc_module":4,"doc_module_name":9,"category_name":20,"show_sort_weight":21,"slug":22},4,"Exam",70,"exam",{"id":24,"doc_module":4,"doc_module_name":9,"category_name":25,"show_sort_weight":26,"slug":27},5,"Comic",60,"comic",{"id":29,"doc_module":4,"doc_module_name":9,"category_name":30,"show_sort_weight":31,"slug":32},6,"Technology",50,"technology",{"id":34,"doc_module":4,"doc_module_name":9,"category_name":35,"show_sort_weight":36,"slug":37},7,"Healthcare",40,"healthcare",{"id":39,"doc_module":4,"doc_module_name":9,"category_name":40,"show_sort_weight":41,"slug":42},8,"Research & Report",30,"research-report",{"id":44,"doc_module":4,"doc_module_name":9,"category_name":45,"show_sort_weight":46,"slug":47},9,"Religion & Spirituality",20,"religion-spirituality",{"id":46,"doc_module":4,"doc_module_name":9,"category_name":49,"show_sort_weight":46,"slug":50},"World Cup","world-cup",{"id":52,"doc_module":4,"doc_module_name":9,"category_name":53,"show_sort_weight":52,"slug":54},10,"Lifestyle","lifestyle",{"id":56,"doc_module":4,"doc_module_name":9,"category_name":57,"show_sort_weight":24,"slug":58},19,"General","general",{"code":4,"msg":60,"data":61},"ok",{"site_id":62,"language":63,"slug":64,"title":65,"keywords":66,"description":67,"schema_data":68,"social_meta":123,"head_meta":125,"extra_data":127,"updated_unix":129},105,"en","thinking-with-images-for-multimodal-reasoning-foundations-methods-and-future-frontiers","Thinking with Images for Multimodal Reasoning - Foundations, Methods, and Future Frontiers","","Recent progress in multimodal reasoning has been driven by textual Chain-of-Thought, where models reason in language while treating vision as a fixed starting context. This creates a semantic gap between rich perceptual signals and discrete symbolic thinking. Humans instead use vision as a dynamic mental sketchpad, and the same shift is emerging in AI. The survey formalizes this paradigm shift toward models that think with images, leveraging visual information as intermediate cognitive steps and assessing the resulting evaluations, benchmarks, and transformative applications across three stages.",{"@graph":69,"@context":122},[70,84,105],{"@type":71,"itemListElement":72},"BreadcrumbList",[73,77,79,82],{"item":74,"name":75,"@type":76,"position":8},"https://docshare.wps.com","Home","ListItem",{"item":78,"name":9,"@type":76,"position":14},"https://docshare.wps.com/document/",{"item":80,"name":40,"@type":76,"position":81},"https://docshare.wps.com/document/research-report/",3,{"item":83,"name":65,"@type":76,"position":19},"https://docshare.wps.com/document/thinking-with-images-for-multimodal-reasoning-foundations-methods-and-future-frontiers/139166/",{"url":83,"name":65,"@type":85,"image":86,"author":91,"headline":65,"publisher":94,"fileFormat":97,"inLanguage":63,"description":67,"dateModified":98,"datePublished":99,"encodingFormat":97,"isAccessibleForFree":100,"interactionStatistic":101},"DigitalDocument",{"url":87,"@type":88,"width":89,"height":90},"https://docshare.wps.com/thumbnails/thinking-with-images-for-multimodal-reasoning-foundations-methods-and-future-frontiers/139166.png","ImageObject",300,407,{"name":92,"@type":93},"Oliver Hayes","Person",{"url":74,"name":95,"@type":96},"DocShare","Organization","application/pdf","2026-10-17","2026-08-23",true,{"@type":102,"interactionType":103,"userInteractionCount":29},"InteractionCounter",{"@type":104},"ViewAction",{"@type":106,"mainEntity":107},"FAQPage",[108,114,118],{"name":109,"@type":110,"acceptedAnswer":111},"Why does textual Chain-of-Thought create a limitation for multimodal reasoning?","Question",{"text":112,"@type":113},"Textual CoT treats vision as a static initial context, which leaves a semantic gap between perceptual information and symbolic reasoning steps. The survey motivates a paradigm where vision contributes dynamically to the reasoning process.","Answer",{"name":115,"@type":110,"acceptedAnswer":116},"What are the three stages of the “Thinking with Images” paradigm?",{"text":117,"@type":113},"The survey organizes the evolution into three stages: Stage 1 tool-driven visual exploration, Stage 2 programmatic visual manipulation, and Stage 3 intrinsic visual imagination. Each stage increases cognitive autonomy in how visual information is used.",{"name":119,"@type":110,"acceptedAnswer":120},"What does the survey cover regarding evaluation and applications?",{"text":121,"@type":113},"It reviews evaluation benchmarks and implementation frameworks, then analyzes transformative applications across interactive systems, embodied AI/robotics, AI for science, healthcare and medicine, and education and training.","https://schema.org",{"og:url":83,"og:type":124,"og:title":65,"og:site_name":95,"og:description":67},"article",{"robots":126,"canonical":83},"index,follow",{"doc_id":128,"site_id":62},139166,1787505509,{"code":4,"msg":5,"data":131},{"doc_id":128,"user_id":132,"nickname":92,"user_avatar":133,"doc_module":4,"category_id":39,"category_name":40,"doc_title":65,"doc_description":67,"doc_content":134,"file_id":135,"file_url":136,"file_type":137,"file_size":138,"view_count":29,"is_deleted":4,"is_public":8,"is_downloadable":8,"audit_status":8,"page_count":139,"language":140,"language_code":63,"site_id":62,"html_lang":63,"table_of_contents":141,"faqs":142,"seo_title":143,"seo_description":67,"update_tm":129,"read_time":144},687207020761,"https://ap-avatar.wpscdn.com/davatar_155a257f0dc6eb9ab79c44ca47cae57d","arXiv :2506 .239 18v 3 [ cs .CV] 3 Jul 2025  \nThinking with Images for Multimodal Reasoning: Foundations, Methods, and Future Frontiers  \nZhaochen Su 1 , Peng Xia2 , Hangyu Guo 1 , Zhenhua Liu 1 , Yan Ma, Xiaoye Qu, Jiaqi Liu2 , Yanshu Li 1 , Kaide Zeng2 , Zhengyuan Yang3 , Linjie Li3 , Yu Cheng4 , Heng Ji5 , Junxian He 1 , Yi R. (May) Fung 1  \n1The Hong Kong University of Science and Technology  \n2UNC-Chapel Hill, 3Microsoft, 4The Chinese University of Hong Kong, 5UIUC  \nAbstract  \nRecent progress in multimodal reasoning has been significantly advanced by textual Chain-of-Thought (CoT), a paradigm where models conduct reasoning within language. This text-centric approach, however, treats vision as a static, initial context, creating a fundamental semantic gap between rich perceptual data and discrete symbolic thought. Human cognition often transcends language, utilizing vision as a dynamic mental sketchpad. A similar evolution is now unfolding in AI, marking a fundamental paradigm shift from models that merely think about images to those that can truly think with images. This emerging paradigm is characterized by models leveraging visual information as intermediate steps in their thought process, transforming vision from a passive input into a dynamic, manipulable cognitive workspace. In this survey, we chart this evolution of intelligence along a spectrum of increasing cognitive autonomy, which unfolds across three key stages:  \nfrom external tool exploration, through programmatic manipulation, to intrinsic imagination. To structure this rapidly evolving field, our survey makes four key contributions. (1) We establish the foundational principles of the “Thinking with Images” paradigm and its three-stage framework. (2) We provide a comprehensive review of the core methods that characterize each stage of this roadmap. (3) We analyze the critical landscape of evaluation benchmarks and transformative applications. (4) We identify significant challenges and outline promising future directions. Through this structured overview, we aim to offer a clear roadmap for future research towards more powerful and human-aligned multimodal AI.1  \nThe Paradigm Shift in Multimodal Reasoning  \nThink about Images Think with Images  \nThink with Text  \nTextual reasoning over a static visual context.  \nIntelligence Increasing  \nStage 1: Tool-Driven Visual Exploration  \nCommanding a fixed toolkit for visual analysis.  \nStage 2: Programmatic  \nStage 3: Intrinsic Visual Imagination  \nVisual Manipulation  \nCreating bespoke visual  \nInternally generating visual thoughts within a cognitive loop.  \noperations via code generation.  \n1We maintain a real-time GitHub repository tracking progress at: [https://github.com/zhaochen0110/](https://github.com/zhaochen0110/)[ ](https://github.com/zhaochen0110/)Awesome_Think_With_Images.  \nContents  \n1 Introduction 4  \n1.1 The Position of Our Survey .............................. 5  \n2 Foundations ofthe Thinking with Images Paradigm 6  \n2.1 The Cognitive Imperative: Why We Need “Thinking with Images”? ........ 6  \n2.2 How Does It Work? The Three-Stage Evolution ................... 7  \n2.3 What is Thinking with Images? A Formal Definition ................ 8  \n2.4 Unique Challenges .................................. 8  \n3 Stage 1: Tool-Driven Visual Exploration 9  \n3.1 Formulation ...................................... 9  \n3.2 Categories of Visual Tools .............................. 10  \n3.3 Implementation Approaches ............................. 11  \n3.3.1 Prompt-Based Approaches .......................... 11  \n3.3.2 SFT-Based Approaches ............................ 13  \n3.3.3 RL-Based Approaches ............................ 14  \n3.4 Conclusion and Future Frontiers ........................... 14  \n4 Stage 2: Programmatic Visual Manipulation 15  \n4.1 The Programmatic Leap: From Selection to Creation ................ 16  \n4.2 Formulation ...................................... 16  \n4.3 Categories of Composable Operations","cbCaiqGB5MwWMhr5","https://ap.wps.com/l/cbCaiqGB5MwWMhr5","pdf",15860082,57,"English","# Introduction\n## The Position of Our Survey\n# Foundations ofthe Thinking with Images Paradigm\n## The Cognitive Imperative: Why We Need “Thinking with Images”?\n## How Does It Work? The Three-Stage Evolution\n## What is Thinking with Images? A Formal Definition\n## Unique Challenges\n# Stage 1: Tool-Driven Visual Exploration\n## Formulation\n## Categories of Visual Tools\n## Implementation Approaches\n## Conclusion and Future Frontiers\n# Stage 2: Programmatic Visual Manipulation\n## The Programmatic Leap: From Selection to Creation\n## Formulation\n## Categories of Composable Operations\n## Implementation Approaches\n## Conclusion and Future Frontiers\n# Stage 3: Intrinsic Visual Imagination\n## The Architectural Leap: From Execution to Imagination\n## Formulation\n## Categories of Generative Thoughts\n## Implementation Approaches\n## Conclusion and Future Frontiers\n# Evaluations & Frameworks for Thinking with Images\n## Evaluations Benchmarks\n## Implementation Frameworks\n## Conclusion and Future Frontiers\n# Applications\n## Interactive Systems and User Interfaces\n## Embodied AI and Robotics\n## AI for Science\n## Healthcare and Medicine\n## Education and Training\n# Future Directions\n## Computational and Cognitive Efficiency\n## Safe and Trustworthy Visual Cognition\n## Novel Benchmarks and Evaluation Methodologies\n## Thinking with Audio, Video, and the World\n## Open Questions and Outlook\n# Conclusion","[{\"question\":\"Why does textual Chain-of-Thought create a limitation for multimodal reasoning?\",\"answer\":\"Textual CoT treats vision as a static initial context, which leaves a semantic gap between perceptual information and symbolic reasoning steps. The survey motivates a paradigm where vision contributes dynamically to the reasoning process.\"},{\"question\":\"What are the three stages of the “Thinking with Images” paradigm?\",\"answer\":\"The survey organizes the evolution into three stages: Stage 1 tool-driven visual exploration, Stage 2 programmatic visual manipulation, and Stage 3 intrinsic visual imagination. Each stage increases cognitive autonomy in how visual information is used.\"},{\"question\":\"What does the survey cover regarding evaluation and applications?\",\"answer\":\"It reviews evaluation benchmarks and implementation frameworks, then analyzes transformative applications across interactive systems, embodied AI/robotics, AI for science, healthcare and medicine, and education and training.\"}]","Thinking with Images for Multimodal Reasoning - Foundations, Methods, and Future Frontiers | PDF",144]