[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-84112-en":3,"doc-seo-84112-105":29,"detail-sidebar-cat-0-en-105":95},{"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},84112,1099514067438,"River Wang","https://ap-avatar.wpscdn.com/avatar/100002539ee87300030?x-image-process=image/resize,m_fixed,w_180,h_180&k=1780474512215547542",8,"Research & Report","From Foundation to Application: Improving VLA Models in Practice","Vision-language-action (VLA) foundation models show strong promise for generalist robot policies, yet laboratory conditions remain misaligned with real-world deployment. LingBot-VLA 2.0 bridges this gap through three improvements: expanded pretraining for better generalization across tasks and robot embodiments, a larger action space supporting whole-body degrees of freedom beyond dual-arm setups, and predictive dynamics modeling using proxy future prediction with semantic priors and geometric depth cues. Evaluations on GM-100 confirm gains in generalist, long-horizon mobile manipulation.","arXiv :2607 .06403v 1 [ cs .RO] 7 Jul 2026  \nFrom Foundation to Application: Improving VLA Models in Practice  \nWei Wu∗ , Fangjing Wang∗ , Fan Lu, He Sun, Shi Liu, Yunnan Wang, Yibin Yan, Yong Wang, Shuailei Ma, Xinyang Wang, Yibin Liu, Shuai Yang, Tianxiang Zhou, Kejia Zhang, Lei Zhou, Cheng Su,  \nNan Xue, Bin Tan, Han Zhang, Youchao Zhang, Fei Liao, Xing Zhu, Yujun Shen, Kecheng Zheng†  \n∗Equal Contribution †Project Lead  \nDespite recent progress of VLA foundation models, the disparity between laboratory conditions and realworld applications continues to impede their practical implementation. To bridge this gap, we present LingBot-VLA 2.0, which advances LingBot-VLA through improvements in three functional domains.  \n(1) Generalization across tasks and embodiments. Compared to the previous version, we revamp the data processing pipeline and curate around 60,000 hours of data for pretraining, including 50,000 hours of robot trajectories spanning 20 robot configurations and 10,000 hours of egocentric human videos. (2) Expanded action space in addition to dual-arm hardware platforms. In particular, our system accommodates degrees of freedom for the heads, waists, mobile bases, and dexterous hands, thereby empowering the robots to tackle more complex tasks in practical scenarios. (3) Predictive dynamics modeling for improved temporal reasoning. Specifically, we formulate future prediction as a proxy task, facilitated by a video representation model for semantic priors and a depth estimation model for geometric cues. Evaluations on the GM-100 benchmark, conducted in a generalist setting, validate the beneficial impact of these proposed modifications. Furthermore, benefiting from the expanded pretraining data that covers whole-body degrees of freedom, LingBot-VLA-2.0 demonstrates strong cross-embodiment long-horizon mobile manipulation capability across the two robotic platforms.  \nWebsite: [https://technology.robbyant.com/lingbot-vla-v2](https://technology.robbyant.com/lingbot-vla-v2)  \nGithub: [https://github.com/robbyant/lingbot-vla-v2](https://github.com/robbyant/lingbot-vla-v2)  \nCheckpoints: [https://huggingface.co/collections/robbyant/lingbot-vla-v2](https://huggingface.co/collections/robbyant/lingbot-vla-v2)  \n1 Introduction  \nVision-language-action (VLA) models [4–6, 16] have recently emerged as a promising paradigm for building generalist robot policies. A key advantage of this paradigm is that pretrained vision-language models provide rich multimodal alignment and semantic representations, enabling VLA models to better understand complex scenes and generalize across diverse tasks. Beyond such model-level priors, recent advances [5, 31] further show that scaling up robot data in both quantity and diversity can substantially improve the capability of VLA systems. Together, these developments have established VLA as a compelling foundation for robot learning.  \nHowever, despite this rapid progress, a substantial gap remains between laboratory benchmarks and real-world deployment. In practice, robots are expected to operate under broader embodiment diversity, richer action spaces, and more dynamic environments than those considered in many existing VLA settings. First, generalization in practice isnot only about transferring across tasks, but also about handling heterogeneous robot configurations and data sources. Another point is that many real-world platforms involve substantially more degrees of freedom than standard dual-arm manipulation setups, including head movement, waist, mobile-base control, and dexterous hands. Subsequently, realworld execution often requires anticipating future scene evolution and action consequences, rather than reacting only to current observations. These challenges collectively limit the practical utility of current VLA foundation models.  \n(a) Framework of LingBot-VLA 2.0  \n(b) Dual-Query Distillation  \nFigure 1. Overview of LingBot-VLA 2.0 . We revamp the data processing pipeline and curate","cbCailEDRAkTTZQL","https://ap.wps.com/l/cbCailEDRAkTTZQL","pdf",8211422,1,20,"English","en",105,"# Introduction\n## LingBot-VLA 2.0 framework\n## Data scaling for generalization\n## Expanded action space for real-world embodiments\n## Predictive dynamics modeling","[{\"question\":\"What problem does LingBot-VLA 2.0 address in real-world deployment?\",\"answer\":\"It targets the persistent gap between laboratory benchmarks and practical use, where robots face broader embodiment diversity, richer action spaces, and more dynamic environments.\"},{\"question\":\"How does LingBot-VLA 2.0 improve generalization across tasks and embodiments?\",\"answer\":\"It revamps the data processing pipeline and curates around 60,000 hours of pretraining data, including robot trajectories across multiple configurations and egocentric human videos.\"},{\"question\":\"What enables LingBot-VLA 2.0 to handle more complex actions than dual-arm setups?\",\"answer\":\"It expands the action space by supporting degrees of freedom for the head, waist, mobile base, and dexterous hands.\"},{\"question\":\"How does the model improve temporal reasoning for future execution?\",\"answer\":\"It formulates future prediction as a proxy task, combining a video representation model for semantic priors with a depth estimation model for geometric cues.\"}]",1784192923,50,{"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":90,"head_meta":92,"extra_data":94,"updated_unix":27},"from-foundation-to-application-improving-vla-models-in-practice","",{"@graph":35,"@context":89},[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/from-foundation-to-application-improving-vla-models-in-practice/84112/",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",[71,77,81,85],{"name":72,"@type":73,"acceptedAnswer":74},"What problem does LingBot-VLA 2.0 address in real-world deployment?","Question",{"text":75,"@type":76},"It targets the persistent gap between laboratory benchmarks and practical use, where robots face broader embodiment diversity, richer action spaces, and more dynamic environments.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How does LingBot-VLA 2.0 improve generalization across tasks and embodiments?",{"text":80,"@type":76},"It revamps the data processing pipeline and curates around 60,000 hours of pretraining data, including robot trajectories across multiple configurations and egocentric human videos.",{"name":82,"@type":73,"acceptedAnswer":83},"What enables LingBot-VLA 2.0 to handle more complex actions than dual-arm setups?",{"text":84,"@type":76},"It expands the action space by supporting degrees of freedom for the head, waist, mobile base, and dexterous hands.",{"name":86,"@type":73,"acceptedAnswer":87},"How does the model improve temporal reasoning for future execution?",{"text":88,"@type":76},"It formulates future prediction as a proxy task, combining a video representation model for semantic priors with a depth estimation model for geometric cues.","https://schema.org",{"og:url":51,"og:type":91,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":93,"canonical":51},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":96},[97,101,105,109,114,118,123,126,130,133,137],{"id":20,"doc_module":4,"doc_module_name":45,"category_name":98,"show_sort_weight":99,"slug":100},"Story & Novel",90,"story-novel",{"id":46,"doc_module":4,"doc_module_name":45,"category_name":102,"show_sort_weight":103,"slug":104},"Literature",80,"literature",{"id":52,"doc_module":4,"doc_module_name":45,"category_name":106,"show_sort_weight":107,"slug":108},"Exam",70,"exam",{"id":110,"doc_module":4,"doc_module_name":45,"category_name":111,"show_sort_weight":112,"slug":113},5,"Comic",60,"comic",{"id":115,"doc_module":4,"doc_module_name":45,"category_name":116,"show_sort_weight":28,"slug":117},6,"Technology","technology",{"id":119,"doc_module":4,"doc_module_name":45,"category_name":120,"show_sort_weight":121,"slug":122},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":45,"category_name":12,"show_sort_weight":124,"slug":125},30,"research-report",{"id":127,"doc_module":4,"doc_module_name":45,"category_name":128,"show_sort_weight":21,"slug":129},9,"Religion & Spirituality","religion-spirituality",{"id":21,"doc_module":4,"doc_module_name":45,"category_name":131,"show_sort_weight":21,"slug":132},"World Cup","world-cup",{"id":134,"doc_module":4,"doc_module_name":45,"category_name":135,"show_sort_weight":134,"slug":136},10,"Lifestyle","lifestyle",{"id":138,"doc_module":4,"doc_module_name":45,"category_name":139,"show_sort_weight":110,"slug":140},19,"General","general"]