Group 1 - The core viewpoint of the articles highlights the recent appointment of Yao Shunyu as Tencent's Chief AI Scientist and his focus on developing models and products in the AI sector [1][2] - Tencent is positioned as a company with a strong consumer-oriented (to C) gene, emphasizing the importance of delivering value to users through AI in various scenarios [1] - Yao Shunyu's background includes prestigious education from Tsinghua University and Princeton University, and he is recognized as a leading figure in AI, particularly in the field of large models and intelligent agents [2] Group 2 - The company is considering how to effectively serve its own needs in the B2B sector, noting that larger companies have advantages over startups in terms of established scenarios and ecosystems [2] - Yao Shunyu's contributions to the AI field include the development of the ReAct framework and the "Tree of Thoughts" theory, which are foundational to modern large model reasoning and action coordination [2]

0:00 2025年12月，27岁的姚顺雨正式加入腾讯，担任首席AI科学家，并兼任AI Infra部、大语言模型部负责 人，向总裁刘炽平和技术工程事业群总裁卢山双线汇报。 姚顺雨是AI领域明星研究员。公开资料显示，姚顺雨毕业于清华大学姚班，后就读于普林斯顿大学， 2024年，姚顺雨加入OpenAI。目前，姚顺雨是人工智能领域，特别是大模型"智能体（Agent）"领域的 领军人物，提出了著名的ReAct框架与"思维树（Tree of Thoughts）"理论，奠定了现代大模型推理与行 动协同的技术基石。 他提到，腾讯是一家to C基因很强的公司，会更思考如何让AI给用户提供更多价值，在丰富的场景里落 地AI价值。 B端方面，团队会思考如何先将自身服务好，且大公司做2B业务较创业公司已具备场景和生态优势。 21世纪经济报道记者孔海丽 北京报道 1月10日，在由清华大学基础模型北京市重点实验室、智谱AI发起的AGI-Next前沿峰会上，27岁的姚顺 雨履新腾讯首席AI科学家后首次公开发声，他透露，最近在忙"模型和产品的开发"，也首次谈及了对AI 行业分化趋势的观察。 0:00 ...

据介绍，探迹B2C智能体是基于电商与零售垂直大模型ShopGLM，并融合了DeepSeek、豆包等先进模 型能力，为电商与零售企业提供客服、营销、运营、私域的全流程Agent平台。目前，探迹B2C智能体 已服务超过10000家品牌客户，其中包括妙可蓝多、晨光、格力等知名品牌。 作为专注于数字生产力的大模型智能体平台，探迹科技已构建起覆盖AI Agent全链路的自主技术体系。 公司自主研发的"太擎"企业级大模型智能体开发平台与"旷湖"数据云底座，形成了从智能体开发、数据 治理到场景应用的完整闭环，为商业化落地提供了坚实支撑。 新浪科技讯 1月6日下午消息，探迹B2C智能体近日以年度超10万亿Token消耗量的表现成功跻身"万亿 Tokens俱乐部"。官方披露数据称，探迹B2C智能体日均Token消耗量超500亿，并以每月近20%的速度持 续增长，处于ToB行业Top10之列。这一数据背后，是探迹B2C智能体在电商与零售领域的深度渗透与 规模化应用。 新浪科技讯 1月6日下午消息，探迹B2C智能体近日以年度超10万亿Token消耗量的表现成功跻身"万亿 Tokens俱乐部"。官方披露数据称，探迹B2C智能体日均T ...

来源：科技日报 科技日报记者 罗云鹏 刘侠 12月20日，2025中国多智能体应用大会在四川省成都市举行，与会嘉宾围绕人才培育、技术突破、产业融合与生态共建等核心议题展开交流，共话多智能体 技术生态构建和应用落地的行业新机遇。 中国计算机学会（CCF）理事、中国人工智能学会常务理事、CCF人工智能与模式识别专委会主任梁吉业说，随着大模型的飞跃发展，智能体已然成为定义 下一代软件甚至是操作系统的趋势，今年也被公认为是大模型智能体的元年。 腾讯AI平台部总经理杨巍介绍，腾讯自2019年推出腾讯开悟平台以来，始终致力于通过开放算法、算力和场景资源，推动人工智能科研生态建设，以复杂 游戏为载体模拟真实世界难题，为全球研究者和学子提供高仿真、低成本的实验环境。 大会现场，2026腾讯开悟人工智能全球公开赛与中国大学生服务外包创新创业大赛合作计划发布。双方将共同设计更加贴近产业、更有挑战性的赛题，引导 青年人才关注真实的产业需求，共同孵化既懂技术又能实战的复合型专业人才。 据悉，赛事引入腾讯内部课程体系与技术生态。未来，腾讯青科实训营将成为2026年腾讯开悟人工智能全球公开赛官方线上学习 平台，为参赛学子提供覆 盖全赛程 ...

证券日报网讯 天源迪科11月25日在互动平台回答投资者提问时表示，公司结合行业痛点，携手行业客 户，致力于智能化场景应用。目前，公司有自己的大模型智能体平台，运用多智能体协同技术，打造出 自适应、自主决策、智能交互的行业级解决方案，已应用在运营商、金融等行业；在采购供应链智能化 方面，与众多央国企客户共创智能化应用和解决方案，持续升级核心产品。 （编辑 袁冠琳） ...

Core Viewpoint - The company has developed a comprehensive intelligent product matrix centered around the "Jingbeifang Large Model & Intelligent Agent Series," focusing on four key scenarios: intelligent operation and maintenance, intelligent report generation, intelligent anti-fraud, and intelligent compliance [1] Group 1: Product Development - The company has built a large model intelligent agent platform that is compatible with hardware from Baidu Kunlun and Muxi, and integrates with foundational large models such as Alibaba Tongyi Qianwen, DeepSeek, and Wenxin Yiyan [1] - The products support modular embedding into business systems, leveraging natural language processing, time series analysis, and semantic understanding technologies to enhance business automation and intelligence [1] Group 2: Market Application - The intelligent products have already been implemented in various customer scenarios, demonstrating their effectiveness in meeting the digital transformation needs of industries [1] - The company provides private deployment and model optimization services in a fully domestic environment, addressing the core demands of the Xinchuang (信创) scenario [1]

Core Insights - The release of "Renhe Model 2.0" signifies a major leap for the company from pilot applications of AI technology to a comprehensive promotion of "governance productivity" in smart city management [1][6] - The event highlighted the integration of AI agents in various government sectors, showcasing the practical applications of large model technology in urban governance [1][2] Group 1: Product and Technology Development - "Renhe Model 2.0" utilizes a "data-algorithm-application" matrix to transform business operations from "mouse operations" to "natural language interactions," addressing common pain points in government data applications [2][4] - The technology framework supports rapid deployment, allowing for industry model training within 15 days and application launch within 5 days, enhancing operational efficiency [4] Group 2: Industry Applications - The intelligent agents have been implemented in cities like Beijing, Shenzhen, and Chongqing, effectively converting creative potential into tangible governance productivity [3] - Specific applications include efficient regulation of construction waste, a closed-loop supervision model for administrative law enforcement, and proactive safety systems for urban infrastructure [2][3] Group 3: Collaborative Solutions - The company announced four joint solutions with industry leaders like Huawei and Baidu, marking a new phase in smart city ecosystem collaboration [5][6] - These solutions aim to enhance urban governance capabilities, including a comprehensive monitoring system for low-altitude environments and customizable digital government employees [5][6] Group 4: Strategic Implications - The launch of "Renhe Model 2.0" represents a pivotal moment in the company's evolution, integrating extensive industry knowledge with cutting-edge AI technology to drive business model innovation [6] - The core product, the government AI agent, is expected to provide stable subscription revenue, enhancing the company's profitability and market position in the smart city sector [6]

Core Insights - The article discusses the development of LiveTradeBench, a platform designed to evaluate large language models (LLMs) in real-time trading scenarios, marking a shift from static assessments to dynamic decision-making in financial markets [3][11][34] Group 1: Introduction to LiveTradeBench - LiveTradeBench is initiated by a research team from the University of Illinois Urbana-Champaign, focusing on assessing LLMs' capabilities in real-world trading environments [2] - The platform aims to test LLMs' perception, reasoning, and decision-making abilities through real market dynamics, moving beyond traditional static benchmarks [3][8] Group 2: Key Innovations - LiveTradeBench introduces three core innovations: continuous decision-making, portfolio management, and live trading evaluation, which differentiate it from previous benchmarks [12] - The platform connects directly to real-time stock and prediction market data, eliminating information leakage and allowing for genuine market interaction [15] Group 3: Investment Decision Modeling - The investment decision-making process in LiveTradeBench is modeled as a partially observable Markov decision process (POMDP), requiring LLMs to infer and act based on limited information [19] - The model receives observations that include position information, market prices, and news, enabling it to make informed asset allocation decisions [20][21] Group 4: Performance Evaluation - A 50-day real-world test was conducted on 21 mainstream LLMs, revealing that static reasoning does not equate to effective dynamic decision-making in complex market environments [30] - The results indicated a lack of correlation between high scores in static assessments and actual market performance, highlighting the need for a redefinition of "intelligence" in LLMs [31] Group 5: Future Directions - LiveTradeBench opens new dimensions for evaluating intelligent agents, emphasizing the importance of environmental feedback and continuous decision-making in future AI developments [34] - The platform encourages further research and collaboration in the field of large model agents, inviting students and researchers to engage with ongoing projects [36]

Core Viewpoint - Digital Zhengtong (300075.SZ) is set to hold a conference on October 30, 2025, focusing on the launch of the "Renhe Large Model 2.0" industry AI agent, emphasizing the integration of large model technology in smart city governance [1] Group 1: AI Agent Capabilities - The "Renhe Large Model 2.0" industry AI agent includes three foundational AI agents: Intelligent Query Agent, Intelligent Report Agent, and Wukong Brain Agent [1] - The Intelligent Query Agent leverages over 20 years of management experience from more than 5,000 projects, integrating production data and enhancing statistical results through a comprehensive indicator and rule library [1] - The Intelligent Report Agent transforms data into deep insights and decision-making value, creating a knowledge base that covers data indicators, root cause analysis, and management recommendations [2] Group 2: Urban Governance Applications - The Wukong Brain Agent consolidates experience from over 300 city brain projects, enabling intelligent control of urban management through natural language interaction and data insights [2] - The "Renhe Large Model 2.0" industry AI agent drives urban governance from reactive measures to proactive warnings, enhancing efficiency in various sectors such as construction waste management and administrative law enforcement [2] - The initiative aims to develop a proactive protection system for urban lifelines, enabling efficient hazard detection and pipeline service life warnings [2]

Core Insights - The 2025 Cloud Habitat Conference will be held from September 24 to 26, showcasing AI software products, new models, and infrastructure hardware, with a focus on "cloud intelligence and carbon-silicon symbiosis" [1][2] - Alibaba Cloud will present capabilities in large-scale clusters, distributed training, and inference acceleration, including a high-density AI server supporting 144 computing nodes [1][2] - Over a hundred companies will demonstrate AI applications, with more than 200 Agent applications and 300 AI terminal products on display [1] Software Sector Developments - Companies in the Alibaba Cloud ecosystem are actively developing intelligent agent solutions across various sectors, including finance, government, and healthcare [2][3] - Tianfeng Securities reports that Alibaba Cloud has partnered with 18 companies to release multiple intelligent agents covering areas such as credit risk control and intelligent customer service [3] - New Zhi Software has launched various industry-specific robots, focusing on applications in finance, legal services, and automotive sectors [3][4] Hardware Sector Progress - Alibaba's capital expenditure on AI and cloud infrastructure reached 38.6 billion yuan, a 220% year-on-year increase, indicating significant investment in AI capabilities [6] - Companies in the hardware ecosystem are advancing customer onboarding and product delivery, with Chipone collaborating with RISC-V leaders to provide customized chip services [6][7] - Data center developments are ongoing, with companies like Data Port and Runze Technology expanding their infrastructure to meet national data center layout strategies [7]