[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-84903-en":3,"doc-seo-84903-105":29,"detail-sidebar-cat-0-en-105":91},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":20,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":13,"seo_description":14,"update_tm":27,"read_time":28},84903,8796095461610,"Oliver","https://ap-avatar.wpscdn.com/davatar_276721f389ce27ea32af1340a28f341c",6,"Technology","MoWorld: A Flash World Model","MoWorld presents a cost-effective, high-performance Flash World Model designed to make world modeling practical for real-world autonomy. The framework spans end-to-end pipeline stages: scalable data generation, pre-training, distillation, and efficient inference to support up to 50 FPS interactive performance with cinematic visual quality. MoWorld jointly optimizes capability and cost across development, using a scalable 3D-native data engine for geometrically consistent training data and curriculum cross-frame pre-training. An efficient denoising-step distillation reduces diffusion training cost, while mixed-precision parallel inference enables low-cost real-time deployment on NPU devices. Evaluations show leading performance with 30%–50% lower average inference cost than existing models. Applications include video style transfer, video editing, point cloud reconstruction, and Gaussian splatting.","arXiv :2607 .062 16v 1 [ cs .CV] 7 Jul 2026  \nMoWorld: A Flash World Model  \nMoWorld Team, Moxin Technology  \nAbstract  \nThe future of World Models depends not only on scaling model capability, but also on scaling practicality and inference efficiency. High-frame-rate inference enables responsive perception, planning, and control in real-world autonomous systems. To this end, we present MoWorld, a cost-effective yet high-performance Flash World Model with an end-to-end framework spanning data generation, pre-training, distillation, and efficient inference, enabling up to 50 FPS real-time interaction with cinematic visual quality without the need of high-end GPUs. To enable large-scale realworld deployment, MoWorld jointly optimizes model capability and cost throughout the entire development pipeline. Specifically, unlike existing approaches that primarily rely on large-scale video corpora, MoWorld is built upon a scalable 3D-native data engine accumulated from our large-scale 3D vision and generative modeling pipeline, enabling the efficient construction of geometrically consistent training data across diverse real-world and synthetic environments. Based on this foundation, a curriculum cross-frame pre-training strategy for stable and scalable World Model learning, an efficient denoising-step distillation algorithm to reduce diffusion training cost, and a mixed-precision parallel inference framework for low-cost real-time deployment. MoWorld is the first real-time interactive World Model built on the Neural Processing Unit (NPU) and can achieves up to 50 FPS in such the devices, enabling practical and efficient deployment at scale. Comprehensive evaluations demonstrate that MoWorld achieves leading performance; notably, its average inference cost is only 30%-50% of that of existing World Models, providing a practical foundation for large-scale real-world applications of World Models. We also demonstrate diverse applications of MoWorld, include Video Style Transfer, Video Editing, Point Cloud Reconstruction, Gaussian Splatting and more.  \nProject Page: [https://moxin-tech.github.io/moworld/](https://moxin-tech.github.io/moworld/)  \n1 Introduction  \nFoundation Models [1–14] have rapidly transformed the landscape of artificial intelligence, driving an unprecedented leap toward more general and capable intelligent systems. Following the remarkable success of Large Language Models (LLMs) [8, 15–17] and Multimodal Large Language Models (MLLMs) [18–23], the next frontier is shifting from understanding and generating content to modeling the physical world itself. In this context, World Models are emerging as a new generation of foundation models that enable intelligent agents to perceive, reason about, and interact with dynamic environments [24–37] . Unlike conventional foundation models, which focus primarily on understanding and generating observations, World Models shift the modeling objective from observations to the world itself. Their goal is not merely to synthesize realistic images or videos, but to learn the underlying dynamics that govern how the physical world evolves under environmental changes and embodied agent actions. Equipped with such an internal model of the world,  \nFigure 1 Overview of MoWorld applications across diverse downstream  \ntasks.  \nintelligent agents can predict future states, simulate interactions, reason over alternative outcomes, and plan long-horizon behaviors, making World Models a fundamental element of general embodied intelligence.  \nRather than pursuing ever-larger World Models, we argue that the next stage of the field lies in making them practical. Future World Models must jointly optimize model capability, computational efficiency, deployment cost, and real-time inference performance, enabling scalable adoption beyond research laboratories into realworld intelligent systems. In particular, high-frame-rate inference is essential for latency-sensitive closed-loop applications such as embodied ","cbCaii9Q1kAQO5ek","https://ap.wps.com/l/cbCaii9Q1kAQO5ek","pdf",41334579,1,30,"English","en",105,"# Introduction\n## Flash World Models for Practical Real-Time Interaction\n## MoWorld End-to-End Data–Algorithm–System–Hardware Co-Design\n## Four Pillars Across the World Model Lifecycle","[{\"question\":\"What problem does MoWorld address for World Models?\",\"answer\":\"MoWorld targets practicality by scaling computational efficiency and inference performance so world modeling can run in real-world, closed-loop autonomous systems with low latency.\"},{\"question\":\"How does MoWorld achieve real-time interaction speed?\",\"answer\":\"It uses an end-to-end framework with efficient inference that enables up to 50 FPS without requiring high-end GPUs, and defines Flash World Models as sustaining 30 FPS or higher.\"},{\"question\":\"What techniques reduce MoWorld’s training and inference costs?\",\"answer\":\"MoWorld employs curriculum cross-frame pre-training for stable learning, an efficient denoising-step distillation algorithm to reduce diffusion training cost, and mixed-precision parallel inference for low-cost real-time deployment.\"}]",1784199266,76,{"code":4,"msg":30,"data":31},"ok",{"site_id":24,"language":23,"slug":32,"title":13,"keywords":33,"description":14,"schema_data":34,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":27},"moworld-a-flash-world-model","",{"@graph":35,"@context":85},[36,53,68],{"@type":37,"itemListElement":38},"BreadcrumbList",[39,43,47,50],{"item":40,"name":41,"@type":42,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":44,"name":45,"@type":42,"position":46},"https://docshare.wps.com/document/","Document",2,{"item":48,"name":12,"@type":42,"position":49},"https://docshare.wps.com/document/technology/",3,{"item":51,"name":13,"@type":42,"position":52},"https://docshare.wps.com/document/moworld-a-flash-world-model/84903/",4,{"url":51,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":23,"description":14,"dateModified":61,"datePublished":62,"encodingFormat":60,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":40,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-07-17","2026-07-16",true,{"@type":65,"interactionType":66,"userInteractionCount":20},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"What problem does MoWorld address for World Models?","Question",{"text":75,"@type":76},"MoWorld targets practicality by scaling computational efficiency and inference performance so world modeling can run in real-world, closed-loop autonomous systems with low latency.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How does MoWorld achieve real-time interaction speed?",{"text":80,"@type":76},"It uses an end-to-end framework with efficient inference that enables up to 50 FPS without requiring high-end GPUs, and defines Flash World Models as sustaining 30 FPS or higher.",{"name":82,"@type":73,"acceptedAnswer":83},"What techniques reduce MoWorld’s training and inference costs?",{"text":84,"@type":76},"MoWorld employs curriculum cross-frame pre-training for stable learning, an efficient denoising-step distillation algorithm to reduce diffusion training cost, and mixed-precision parallel inference for low-cost real-time deployment.","https://schema.org",{"og:url":51,"og:type":87,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":89,"canonical":51},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":92},[93,97,101,105,110,113,118,122,127,130,134],{"id":20,"doc_module":4,"doc_module_name":45,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":46,"doc_module":4,"doc_module_name":45,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":52,"doc_module":4,"doc_module_name":45,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":45,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":11,"doc_module":4,"doc_module_name":45,"category_name":12,"show_sort_weight":111,"slug":112},50,"technology",{"id":114,"doc_module":4,"doc_module_name":45,"category_name":115,"show_sort_weight":116,"slug":117},7,"Healthcare",40,"healthcare",{"id":119,"doc_module":4,"doc_module_name":45,"category_name":120,"show_sort_weight":21,"slug":121},8,"Research & Report","research-report",{"id":123,"doc_module":4,"doc_module_name":45,"category_name":124,"show_sort_weight":125,"slug":126},9,"Religion & Spirituality",20,"religion-spirituality",{"id":125,"doc_module":4,"doc_module_name":45,"category_name":128,"show_sort_weight":125,"slug":129},"World Cup","world-cup",{"id":131,"doc_module":4,"doc_module_name":45,"category_name":132,"show_sort_weight":131,"slug":133},10,"Lifestyle","lifestyle",{"id":135,"doc_module":4,"doc_module_name":45,"category_name":136,"show_sort_weight":106,"slug":137},19,"General","general"]