新浪科技讯 1月7日晚间消息，今日北京人形机器人创新中心，携"具身天工2.0"、"具身天工Ultra"等多 款机器人亮相CES 2026。 展会现场，"具身天工2.0 "展示全自主分拣零部件并与现场观众互动，让观众沉浸式体验了具身智能作 业"快、准、狠"的特点。在现场展示中，基于北京人形"慧思开物"平台已发布并开源的跨本体VLA模 型，"具身天工"机器人可以自主连贯高效完成抓取、分类、码放等全流程操作，可以灵活应对包括物体 姿态变化、传送带外背景变化、空间位置变化等影响，展现了极强的泛化性，具备三大突出优势：一是 快，北京人形首创的UVMC（多模态视动统一表征）技术，搭建起视觉与动作的映射桥梁，让机器人将 看到的画面瞬间转化为身体的本能反应，像人类条件反射般自然做出正确的应对动作，从而游刃有余地 应对分拣场景中的突发情况；二是准，具备大于60赫兹的高频率控制能力，可以将视觉信息实时转化为 流畅、准确的动作指令，并且完成高速的动态精准抓取，真正打破"看到"与"做到"的界限；三是狠，具 备出色的双臂协作能力，在右臂错过工件抓取时，左臂能及时协同完成抓取，确保分拣任务的连续性与 可靠性。 据悉，目前，北京人形已经与 ...

● 本报记者 倪铭娅 熊彦莎 连润 开源"具身天工"平台与"慧思开物"框架，推动行业共享底层技术；建设头部产线与"机器人生产机器 人"柔性模式，助推量产时代加速到来…… 走进北京经开区，机器人产业的蓬勃朝气扑面而来。作为北京机器人产业发展高地，这里正凝聚政策赋 能、金融支撑、生态协同三重动能，构建起从核心零部件研发到多场景商业化落地的全产业链体系，形 成机器人跑进千行百业的独特生态，为我国人形机器人产业高质量发展打造"亦庄样板"。 政策精准赋能 "欢迎来到创新展厅，接下来由我带您探索人形机器人的应用世界。"踏入北京人形机器人创新中心展 厅，一台人形机器人缓缓迈步上前，声音温和而清晰。 记者了解到，这位"讲解员"是升级版的"具身天工2.0"。跟随它的指引来到核心展区，只见它在进行讲 解的同时，通过"慧思开物"平台向其他机器人下达指令，物料搬运、倒料、分拣、打包等一系列任务被 其他机器人自主完成。 眼前所见，依托的是北京人形机器人创新中心12月11日发布的全国首个全自主无人化人形机器人导览解 决方案。该方案以"慧思开物"平台打造的"感知—决策—执行"全链路闭环技术体系为核心，依托AI大模 型驱动的具身"大脑"、数 ...

本报讯 （记者贾丽）12月18日，北京人形机器人创新中心（以下简称"北京人形"）正式开源国内首个 且唯一通过具身智能国标测试的具身VLA大模型XR-1以及配套的数据基础RoboMIND2.0、ArtVIP最新 版。基于上述开源成果，能够推动具身智能行业追寻最本质需求，让机器人真正在各类应用场景下能干 活并且会干活，推动国内具身智能行业迈向"全自主、更好用"的新阶段。 聚焦"让机器人能干活、会干活"的核心诉求，北京人形打造了通用机器人平台"具身天工"、通用具身智 能平台"慧思开物"。围绕"具身天工"平台，北京人形已发布具身天工2.0、天轶2.0等多类型通用本体， 为人形机器人干活奠定物理本体基础。而具身大脑和小脑的协同是人形机器人干活的另一前提，目前围 绕"慧思开物"，北京人形已开源WoW（我悟）世界模型和Pelican-VL（天鹕）等具身大脑相关成果。 本次开源的系列内容，是面向具身小脑能力的VLA模型XR-1，以及为XR-1等模型提供数据训练支持的 RoboMIND2.0和ArtVIP。 让机器人拥有"本能" 当前，具身智能行业面临着一个核心痛点：AI技术虽能实现文本创作、视频生成等虚拟场景应用，但 机器人 ...

- 编者按 "十四五"即将圆满收官，"十五五"即将迎来开局之年。2025年，我国经济攻坚克难、稳健前行，展现出 强大韧性和活力。科技创新取得丰硕成果，全国统一大市场建设向纵深推进。 回望来路，砥砺前行。本报今起推出"年终报道"专栏，回首2025年经济运行新变化、新亮点，前瞻2026 年政策新动向、发展新机遇。 2025年是"十四五"规划收官之年。中国经济以韧性破局，向新向优迈进，勾勒高质量发展的清晰路径。 政策护航精准有力、科技创新持续赋能、市场信心稳步提升，共同铸就了经济"进"的核心密码。 2026年是"十五五"规划开局之年。随着增量政策持续发力扩大内需基本盘、新质生产力加速壮大点燃发 展新引擎，中国经济必将不断巩固稳中向好态势，在质的有效提升和量的合理增长中，书写更加精彩的 发展篇章，为全球经济增长持续贡献中国力量。 看总量：有望迈上新台阶 "我临时决定12月底去滑雪，工作忙没时间做攻略，靠在线旅行平台的AI助手给我提供了三日游方 案。"在北京工作的刘女士说，平台AI助手给出的旅游攻略特别全，除了帮她安排好日程、特色餐饮， 还提供了不同旅行方式的价位参考，地图编辑模式也非常清楚。 刘女士的例子是今年"人工 ...

2025年是"十四五"规划收官之年。中国经济以韧性破局，向新向优迈进，勾勒高质量发展的清晰路径。 政策护航精准有力、科技创新持续赋能、市场信心稳步提升，共同铸就了经济"进"的核心密码。 2026年是"十五五"规划开局之年。随着增量政策持续发力扩大内需基本盘、新质生产力加速壮大点燃发 展新引擎，中国经济必将不断巩固稳中向好态势，在质的有效提升和量的合理增长中，书写更加精彩的 发展篇章，为全球经济增长持续贡献中国力量。 看总量：有望迈上新台阶 "我临时决定12月底去滑雪，工作忙没时间做攻略，靠在线旅行平台的AI助手给我提供了三日游方 案。"在北京工作的刘女士说，平台AI助手给出的旅游攻略特别全，除了帮她安排好日程、特色餐饮， 还提供了不同旅行方式的价位参考，地图编辑模式也非常清楚。 刘女士的例子是今年"人工智能+消费"场景不断扩展的缩影。截至2025年6月，我国生成式人工智能用户 规模达5.15亿人，较2024年12月增长2.66亿人，成为新消费活力迸发的生动写照。 消费提质，投资增效。12月1日，厦金大桥厦门段主塔以"毫米级"精度圆满封顶，大桥建设从"立塔成 型"转入"悬索成桥"关键施工阶段；12月5日，南宁至 ...

Economic Overview - In 2025, China's economy is expected to reach approximately 140 trillion yuan, having crossed significant thresholds of 110 trillion, 120 trillion, and 130 trillion yuan [2] - The GDP growth rate for the first three quarters of 2025 is reported at 5.2%, an increase of 0.4 percentage points compared to the same period last year [2] - The World Bank and the International Monetary Fund have raised their economic growth forecasts for China, with the IMF projecting a growth rate of 5% for 2025, up by 0.2 percentage points from previous estimates [3] Investment and Infrastructure - Major infrastructure projects, such as the completion of the Xiamen section of the Xiamen-Jinmen Bridge and the full operation of the Nanning-Pingxiang high-speed railway, highlight significant investment efficiency improvements [2] - The issuance of 1.3 trillion yuan in long-term special bonds and the establishment of 500 billion yuan in new policy financial tools are aimed at stimulating investment [3] Technological Innovation - By mid-2025, the user base for generative artificial intelligence in China reached 515 million, reflecting a significant increase of 266 million users from December 2024 [2] - China's innovation index has entered the global top ten for the first time, with substantial growth in the equipment manufacturing and digital industries, both showing a 9.5% increase in value added [5] - The development of new technologies, including artificial intelligence, is driving industry transformation and creating new economic momentum [5][6] Market Confidence - High-tech industries attracted 192.52 billion yuan in foreign investment in the first ten months of 2025, with significant growth in sectors such as e-commerce and medical equipment [7] - The A-share market has seen increased trading activity, with daily transaction volumes frequently exceeding 2 trillion yuan [7] - The implementation of policies promoting consumption, such as the "old-for-new" program, has led to notable sales growth in the home appliance and communication retail sectors [7] Future Outlook - The Central Economic Work Conference emphasizes the need for proactive macroeconomic policies to ensure quality and reasonable growth in 2026, the first year of the 15th Five-Year Plan [4][6] - The focus will be on enhancing the integration of existing and new policies to support economic stability and growth [4]

Group 1 - Beijing Humanoid Robot Innovation Center has launched the open-source embodied intelligence platform "Wisdom Open" SDK in collaboration with UBTECH, marking a new phase in the open-source ecosystem for embodied intelligence [1][2] - The newly introduced open-source embodied world model WoW aims to help robots quickly learn various skills, enhancing their operational capabilities in real-world applications [1][2] - The development of specialized models for material handling has improved robots' generalization recognition abilities, allowing them to efficiently handle various types of cargo boxes [1] Group 2 - The embodied intelligence platform and world model have established a comprehensive data system that supports high-quality training for different robotic bodies and task scenarios [2] - The dual open-source strategy of "platform + model" is seen as a foundational infrastructure for the industry, potentially attracting more developers and fostering a cycle of technological iteration [2] - The commercialization of humanoid robots is accelerating, with Beijing Humanoid's robots already operational in factories, performing tasks such as box handling and transportation [2] Group 3 - Other companies in the industry are also pushing for commercialization, with UBTECH achieving mass application of its Walker S robot in industrial settings [3] - The humanoid robot industry is experiencing rapid development, with open-source ecosystems becoming a competitive advantage for leading companies [3] - The focus is shifting from single-point technological breakthroughs to competitive ecosystems, particularly in standardized industrial applications, which are easier to commercialize [3]

Core Viewpoint - The integration of humanoid robots into factories is progressing, with a focus on enhancing operational efficiency and reducing manual labor risks through advanced robotic solutions [1][2][3]. Group 1: Robot Deployment and Functionality - The Tianyi 2.0 humanoid robot is currently being tested for box handling tasks at the Foton Cummins engine factory, showcasing its ability to lift and categorize boxes weighing between 8 to 12 kilograms [1]. - The deployment of humanoid robots will occur in two phases: initially assisting human workers and later integrating with factory systems for direct task execution [2]. - The Tianyi 2.0 robot is designed to improve its decision-making capabilities and efficiency in repetitive tasks, with current handling times of 1 to 2 minutes per box [2]. Group 2: Industry Collaborations and Orders - Recent collaborations indicate a surge in orders for humanoid robots, with companies like Zhiyuan Robotics announcing significant contracts worth hundreds of millions, including a multi-million yuan order from Longqi Technology [3][6]. - The deployment of the new generation interactive humanoid robot, Zhiyuan Spirit G2, is set to enhance operations in consumer electronics and automotive parts manufacturing [3]. - Companies like Yubiquitous and Xian Gong Intelligent are also entering into large-scale agreements to deploy humanoid robots across various industrial applications, with plans for over 1,000 units in the next few years [5][6]. Group 3: Technological Advancements and Market Trends - The humanoid robot market is in its early commercial phase, with a focus on practical applications in real-world scenarios to refine technology [4]. - The emphasis is on operational capabilities over aesthetic features, with ongoing improvements in interaction latency and task execution speed [5]. - The industry is witnessing a trend towards integrating humanoid robots into logistics, manufacturing, and quality inspection processes, indicating a growing acceptance and reliance on robotic solutions [5][6].

Core Viewpoint - The deployment of embodied intelligent robots, specifically Tianyi 2.0, at the Foton Cummins engine factory marks a significant step in automating material handling processes, aiming to enhance efficiency and reduce risks associated with manual labor [1][3][4]. Group 1: Robot Deployment and Functionality - Tianyi 2.0 is designed to accurately identify the position of material boxes weighing between 8 to 12 kilograms and can lift them to heights of up to 1.7 meters, reducing the risk of injury from manual handling [3][4]. - The implementation of these robots is in two phases: initially as assistants to human workers, followed by integration with the factory's smart systems for direct task execution [4][5]. - The robots are expected to improve in load capacity, operational precision, and task execution efficiency, with current handling times of 1 to 2 minutes per task [4][6]. Group 2: Industry Trends and Collaborations - Recent collaborations in the robotics sector include significant orders for humanoid robots, with companies like Zhiyuan Robotics and Yubiquitous securing multi-million dollar contracts for deployment in various manufacturing scenarios [6][8]. - The deployment of humanoid robots is expanding into sectors such as 3C manufacturing and automotive parts production, with companies like Zhiyuan Robotics and Smart Square announcing plans for large-scale implementations [6][7]. - The industry is witnessing a trend towards integrating humanoid robots into logistics and manufacturing processes, with partnerships formed to enhance operational capabilities across multiple sectors [7][8].

Core Viewpoint - The article highlights the advancements in autonomous navigation technology for humanoid robots, particularly focusing on the "Embodied Tiangong Ultra" robot, which achieved significant milestones in recent competitions without human control, indicating a shift towards full autonomy in robotics [1][2][3]. Group 1: Competition Achievements - The "Embodied Tiangong Ultra" robot won the 100-meter race with a time of 21.50 seconds and secured silver medals in both the 400-meter and 1500-meter events at the recent robotics sports event [1]. - In a previous half-marathon competition, the same robot also claimed victory, showcasing its evolving capabilities in autonomous navigation [2]. Group 2: Technological Advancements - The removal of remote control and the implementation of full autonomous navigation are deemed necessary for the industrial application of robots, allowing them to explore new environments independently [3][8]. - The CTO of the Beijing Humanoid Robot Innovation Center, Tang Jian, emphasized that the current demonstration of autonomous navigation is just a small part of a comprehensive solution, with future developments expected to enhance capabilities further [2][9]. Group 3: Industry Challenges - Achieving full autonomous navigation presents significant challenges due to the complexity of environments robots may encounter, requiring advanced algorithms for real-time mapping and obstacle avoidance [11][12]. - The industry currently lacks a consensus on the capabilities of humanoid robots, with many companies still relying on remote control, which hinders the potential for widespread adoption [19][20]. Group 4: Future Prospects - The expectation is that more companies will adopt full autonomous navigation in future competitions, indicating a potential industry-wide shift towards this technology [17]. - The article suggests that the humanoid robot industry is at a critical juncture, with the need for improved algorithms and models to enhance the generalization capabilities of embodied intelligence [21][22].