[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-85624-en":3,"doc-seo-85624-105":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},85624,4810365810221,"Aurora","https://ap-avatar.wpscdn.com/davatar_155a257f0dc6eb9ab79c44ca47cae57d",8,"Research & Report","Embodied-R1.5: Evolving Physical Intelligence via Embodied Foundation Models","Embodied-R1.5 presents a unified Embodied Foundation Model (EFM) that combines embodied cognition, task planning and correction, and embodied pointing/location into a single architecture for general physical intelligence. Three automated data construction pipelines expand coverage to a large-scale 15B+ token dataset, while a multi-task balanced RL recipe mitigates heterogeneous task conflicts. A PlannerGrounder-Corrector (PGC) closed-loop framework lets one model execute and self-correct long-horizon tasks, reaching SOTA on most embodied VLM benchmarks and strong zero-shot real-robot transfer. The work open-sources model weights, data, training code, and EmbodiedEvalKit.","arXiv :2606 . 11324v2 [ cs .RO] 11 Jul 2026  \nEmbodied-R1.5: Evolving Physical Intelligence via Embodied Foundation Models  \nYifu Yuanv ,\\# 1 , Yaoting Huang1 , Xianze Yao1 , Yutong Li1 , Shuoheng Zhang1 , Linqi Han1 , Pengyi Li1 , Jiangeng Sun1 , Wenting Jia1 , Zhao Zhang1 , Yuhao Liu1 , Ruihao Liao1 , Yucheng Hu1 , Qiyu Wu1 , Yuxiao Li1 , Zibin Dong1 , Fei Ni1 , Yan Zheng1 , Shuyang Gu\\# 2 , Yi Mav ,\\# 1 , Hongyao Tangv ,\\# 1 , Han Hu2 , Jianye Hao\\# 1  \n1Tianjin University, 2Tencent Hunyuan  \nvProject Leader,\\# Corresponding Author (Contact: [yuanyf@tju.edu.cn](yuanyf@tju.edu.cn))  \nWe introduce Embodied-R1.5, a unified Embodied Foundation Model (EFM) that integrates comprehensive embodied reasoning capabilities, spanning embodied cognition, task planning, correction, and pointing, within a single architecture toward general physical intelligence. Leveraging three automated data construction pipelines to significantly expand the data coverage of critical capabilities, we build a large-scale data system of over 15B tokens, and design a multi-task balanced RL recipe to alleviate heterogeneous task conflicts. We further introduce a PlannerGrounder-Corrector (PGC) closed-loop framework that enables a single model to autonomously execute and self-correct over long-horizon tasks. With only 8B parameters, Embodied-R1.5 achieves SOTA on 16 out of 24 embodied VLM benchmarks, surpassing leading models like Gemini-Robotics-ER-1.5 and GPT-5.4 . Benefiting from the internalized embodied capabilities, Embodied-R1.5 can be fine-tuned into a VLA with only a small amount of data, outperforming leading VLA models like 􀀙0.5 across 4 popular manipulation benchmark suites. We further conduct extensive zero-shot real-robot experiments, validating performance in instruction following, affordance grounding, articulated object manipulation, and long-horizon complex tasks, demonstrating strong generalization to the physical world. We open-source model weights, datasets, training code, and EmbodiedEvalKit, an evaluation framework tailored for embodied tasks, to facilitate future research in EFMs.  \n􀂀 Project: [https://embodied-r.github.io/](https://embodied-r.github.io/)  \n Code: [https://github.com/pickxiguapi/Embodied-R1.5](https://github.com/pickxiguapi/Embodied-R1.5)  \n EmbodiedEvalKit: [https://github.com/pickxiguapi/EmbodiedEvalKit](https://github.com/pickxiguapi/EmbodiedEvalKit)  \n Models & Datasets: [https://huggingface.co/collections/Iff Yuan/embodied-r15](https://huggingface.co/collections/Iff Yuan/embodied-r15)  \nLong-horizon Manipulation Instruction Following Tool Affordance Disturbance Spatial Contact-Rich  \nFigure 1 Performance overview of Embodied-R1.5 . Top: Performance across 24 embodied VLM benchmarks (21 main benchmarks + 3 visual trace benchmarks) and 4 robotic manipulation benchmark suites, compared with leading general and embodied models. Bottom: Zero-shot real-robot experiments validating diverse embodied capabilities, including long-horizon manipulation, instruction following, tool affordance, and contact-rich tasks.  \nContents  \n1 Introduction 3  \n2 Unified Embodied Capabilities & Architecture 4  \n2.1 Unified Embodied Capabilities ................................. 5  \n2.2 Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6  \n3 Training Data Construction 6  \n4 Training Strategy 8  \n4.1 Stage 1: Supervised Fine-Tuning ............................... 8  \n4.2 Stage 2: Reinforced Fine-Tuning ................................ 8  \n4.2.1 Multi-Task Balanced RL Recipe ............................ 8  \n4.2.2 Multi-Task Reward Design ............................... 9  \n5 Closed-Loop PGC Autonomy Framework 10  \n6 EmbodiedEvalKit 11  \n7 Experiments 12  \n7. 1 Embodied VLM Benchmarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12  \n7.2 Robotic Manipulation in Simulation .............................. 14  \n7.3 Zero-Shot Manipulation Transfer .............................","cbCaicTP30XxapsL","https://ap.wps.com/l/cbCaicTP30XxapsL","pdf",27920715,1,47,"English","en",105,"# Introduction\n# Unified Embodied Capabilities & Architecture\n## Unified Embodied Capabilities\n## Architecture\n# Training Data Construction\n# Training Strategy\n## Stage 1: Supervised Fine-Tuning\n## Stage 2: Reinforced Fine-Tuning\n# Closed-Loop PGC Autonomy Framework\n# EmbodiedEvalKit\n# Experiments\n## Embodied VLM Benchmarks\n## Robotic Manipulation in Simulation\n## Zero-Shot Manipulation Transfer\n## Long-Horizon Closed-Loop Demonstrations\n## Analysis\n# Related Work\n# Conclusion","[{\"question\":\"What is Embodied-R1.5, and what core capabilities does it integrate?\",\"answer\":\"Embodied-R1.5 is a unified embodied foundation model that integrates embodied cognition, task planning and correction, and embodied pointing/location so the same model can perceive, act, and self-correct in a unified loop.\"},{\"question\":\"How does Embodied-R1.5 scale training data and handle conflicts across tasks?\",\"answer\":\"It uses three automated data construction pipelines to expand data coverage to a 15B+ token system, and applies a multi-task balanced RL recipe to alleviate conflicts among heterogeneous tasks.\"},{\"question\":\"What is the PlannerGrounder-Corrector (PGC) framework used for?\",\"answer\":\"PGC provides a closed-loop autonomy mechanism that enables a single model to autonomously execute long-horizon tasks and self-correct during execution.\"}]",1784205022,118,{"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},"embodied-r15-evolving-physical-intelligence-via-embodied-foundation-models","",{"@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/research-report/",3,{"item":51,"name":13,"@type":42,"position":52},"https://docshare.wps.com/document/embodied-r15-evolving-physical-intelligence-via-embodied-foundation-models/85624/",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 is Embodied-R1.5, and what core capabilities does it integrate?","Question",{"text":75,"@type":76},"Embodied-R1.5 is a unified embodied foundation model that integrates embodied cognition, task planning and correction, and embodied pointing/location so the same model can perceive, act, and self-correct in a unified loop.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How does Embodied-R1.5 scale training data and handle conflicts across tasks?",{"text":80,"@type":76},"It uses three automated data construction pipelines to expand data coverage to a 15B+ token system, and applies a multi-task balanced RL recipe to alleviate conflicts among heterogeneous tasks.",{"name":82,"@type":73,"acceptedAnswer":83},"What is the PlannerGrounder-Corrector (PGC) framework used for?",{"text":84,"@type":76},"PGC provides a closed-loop autonomy mechanism that enables a single model to autonomously execute long-horizon tasks and self-correct during execution.","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},[93,97,101,105,110,115,120,123,128,131,135],{"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":111,"doc_module":4,"doc_module_name":45,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":45,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":45,"category_name":12,"show_sort_weight":121,"slug":122},30,"research-report",{"id":124,"doc_module":4,"doc_module_name":45,"category_name":125,"show_sort_weight":126,"slug":127},9,"Religion & Spirituality",20,"religion-spirituality",{"id":126,"doc_module":4,"doc_module_name":45,"category_name":129,"show_sort_weight":126,"slug":130},"World Cup","world-cup",{"id":132,"doc_module":4,"doc_module_name":45,"category_name":133,"show_sort_weight":132,"slug":134},10,"Lifestyle","lifestyle",{"id":136,"doc_module":4,"doc_module_name":45,"category_name":137,"show_sort_weight":106,"slug":138},19,"General","general"]