Core Insights - Alibaba has launched Qwen3-Max-Thinking, its most advanced reasoning model, which competes effectively with top models like GPT-5.2-Thinking and Gemini 3 Pro across 19 benchmark tests [1][4]. Model Capabilities - Qwen3-Max-Thinking demonstrates strong performance in various benchmarks, achieving scores close to or exceeding those of leading models in categories such as Knowledge, STEM, Reasoning, and Tool Use [3]. - The model features adaptive tool invocation capabilities, allowing it to automatically call upon search engines and code interpreters as needed, enhancing user experience [4][10]. Technical Innovations - The model employs a unique reasoning approach that focuses on efficient use of computational resources, prioritizing intelligent reasoning processes over merely increasing parallel reasoning paths [4][16]. - Qwen3-Max-Thinking integrates a context window of 256k and is believed to have over 1 trillion parameters, similar to its preview version [4]. Pricing and Accessibility - The API for Qwen3-Max-Thinking is available at a competitive rate of 2.5 yuan per million input tokens and 10 yuan per million output tokens, providing good value for users [5][6]. Additional Features - On the same day, Alibaba also released the Qwen3-TTS series, which includes advanced voice synthesis capabilities such as voice cloning and natural language-based voice control [7]. - The model's ability to combine search and code interpretation has been tested, showing its capability to handle complex tasks effectively, although some limitations in search accuracy were noted [10][15]. Industry Context - Alibaba's advancements in AI models reflect a broader trend in the industry towards optimizing computational efficiency and enhancing reasoning capabilities, addressing the constraints of available computational power in AI research [26].

同日晚间，@豆包手机助手 发文称："我们理解外界对于手机助手安全与隐私的关注。豆包手机助手严 格遵循用户授权与合规的原则，仅在用户明确授权的前提下调用必要能力。此外，豆包手机助手在云端 处理用户手机屏幕内容时，严格遵循'不存储、不训练'的原则，数据传输过程加密，使用过程有严格的 保护措施，确保用户隐私安全。（凤凰科技） 阿里深夜发布号称最强千问推理模型，比肩GPT-5.2 1月26日深夜，阿里发布千问旗舰推理模型Qwen3-Max-Thinking。 阿里称，Qwen3-Max-Thinking在多个 关键维度上实现了显著提升，包括事实知识、复杂推理、指令遵循、人类偏好对齐以及智能体能力。在 19项权威基准测试中，其性能可媲美GPT-5.2-Thinking、 Claude-Opus-4.5和Gemini3Pro等顶尖模型。 Qwen3-Max-Thinking引入两项核心创新:1.自适应工具调用能力，可按需调用搜索引擎和代码解释器， 现已上线Qwen Chat;2.测试时扩展技术(Test-Time Scaling)，显著提升推理性能。 21世纪经济报道新质生产力研究院综合报道 早上好，新的一天又开始了。在 ...

早安，艾瑞巴蒂（手动问好），一起来看今天的早报： 【黑马高调早报】第3057期 更多创业内容请访问www.iheima.com Fast Reading 今日头条 | 日 1日 | | | | | --- | --- | --- | --- | | 日通 | 员说 | 中通 | 韵达 | | 1993年创立 | 2000年创立 | 2002年创立 | 1999年创立 | | 创始人 | 创始人 | 创始人 | 创始人 | | 壹脖飞 | 喻渭蛟 | 赖梅松 | 慧腾云 | | 浙江桐庐人 | 浙江桐庐人 | 浙江桐庐人 | 浙江桐庐人 | | 关系网 | 关系网 | 关系网 | 关系网 | | A 中国民营快递 | > 创立圆通前, | > 申通早期加盟 | > 创始人聂腾云, | | 开创者,桐庐 | 其妻张小娟在 | 商,后创立了 | 是申通创始人 | | 帮鼻祖; | 申通做财务; | 中通快递; | 聂腾飞的亲弟 | | | | | 弟: | | A 聂腾飞的妻子 | × 张小娟是申通 | > 与申通陈德军、 | | | 陈小英和大舅 | 陈德军中学同 | 韵达聂腾云是 | > 在哥哥聂腾飞 | | ...

Core Insights - The Chinese internet advertising industry has transitioned from a phase of "incremental traffic acquisition" to a focus on "efficiency extraction from existing resources" [1][2] - The new rule emerging from this differentiation is that technology determines efficiency, efficiency drives revenue, and ecosystem shapes the landscape [2] Group 1: Advertising Revenue Performance - Alibaba's advertising revenue for Q3 2025 reached 789.27 billion yuan, a 10% year-on-year increase and a slight 2.1% quarter-on-quarter growth, maintaining its leading position in the industry [5] - Pinduoduo's online marketing service revenue was 533.48 billion yuan, an 8% year-on-year increase but a 1.2% quarter-on-quarter decline, indicating a strategic focus on long-term ecosystem development [6][7] - Tencent's marketing service revenue grew by 21% to 362 billion yuan, showcasing the effectiveness of its technology-driven approach [7][8] - JD.com's market and marketing revenue increased by 23.7% to 256.88 billion yuan, driven by the growth of its food delivery business and AI-enabled advertising solutions [10] - Bilibili's advertising revenue reached 25.7 billion yuan, with a 23% year-on-year growth, attributed to the maturation of its user base [11] Group 2: Challenges Faced by Mid-Tier Platforms - Baidu's online marketing revenue fell by 18.6% to 153 billion yuan, primarily due to budget cuts from advertisers and the restructuring of search scenarios by AI [12][14] - Zhihu's advertising revenue declined by 26.19%, reflecting a mismatch between its high-knowledge user base and consumer conversion needs [15] - Automotive media service revenue for Che168 dropped by 8.59%, as traditional advertising models struggle against the evolving strategies of new energy vehicle companies [16] Group 3: Industry Trends and Future Outlook - The differentiation in advertising revenue among internet companies in Q3 2025 signifies a reconstruction of internet marketing value logic, focusing on traffic depth, transaction closure, and technology adaptation [16] - Platforms achieving growth typically exhibit a combination of extensive traffic, ecosystem closure, and deep integration of AI technology, while declining platforms often lack these comprehensive advantages [16] - The future of internet marketing will prioritize platforms that can achieve ecosystem synergy, embedded technology, and full-link value delivery, moving away from traditional single-channel models [16]

消息面上，据阿里云公众号消息，1月26日，阿里巴巴正式发布千问旗舰推理模型Qwen3-Max- Thinking。该模型总参数量超万亿(1T)，预训练数据量高达36T Tokens。目前，开发者可在QwenChat上 免费体验Qwen3-Max-Thinking模型，企业可通过阿里云百炼获取新模型API服务，普通用户也可通过千 问PC端和网页端试用模型。此外，千问APP也即将接入新模型，所有用户都可免费体验千问最强模 型。 阿里巴巴-W(09988)涨超2%，截至发稿，涨1.69%，报168港元，成交额30.43亿港元。 ...

智通财经APP获悉，阿里巴巴-W(09988)涨超2%，截至发稿，涨1.69%，报168港元，成交额30.43亿港 元。 消息面上，据阿里云公众号消息，1月26日，阿里巴巴正式发布千问旗舰推理模型Qwen3-Max- Thinking。该模型总参数量超万亿(1T)，预训练数据量高达36T Tokens。目前，开发者可在QwenChat上 免费体验Qwen3-Max-Thinking模型，企业可通过阿里云百炼获取新模型API服务，普通用户也可通过千 问PC端和网页端试用模型。此外，千问APP也即将接入新模型，所有用户都可免费体验千问最强模 型。 ...

【申万互联网传媒】AWS涨价，关注国内云厂投资机会 AWS近期对AI算力服务EC2机器学习容量块进行了约15%的价格上调。 此次涨价反映了上游硬件成本上涨传导，同时确认了下游需求强劲。 对于国内，大厂今年加大AI超级入口投入或加快应用普及，需求端的提价空间逐渐增大。 上游硬件成本上涨传导，叠加部分GPU资源仍供给紧张，国内云厂对部分产品调整价格策略逻辑成立。 但国内云厂竞争更为激烈，后续除观察收入增速，也需关注利润率变化。 国内云厂长期提升利润率关键？ 自研芯片降低成本，全栈AI提升粘性；能做更高利润率的SaaS，出海做大收入空间。 相关标的 【阿里巴巴】国内云厂领军，全栈AI强技术能力，平头哥自研芯片成本优势大； 上游硬件成本上涨传导，叠加部分GPU资源仍供给紧张，国内云厂对部分产品调整价格策略逻辑成 【申万互联网传媒】AWS涨价，关注国内云厂投资机会 AWS近期对AI算力服务EC2机器学习容量块进行了约15%的价格上调。 此次涨价反映了上游硬件成本上涨传导，同时确认了下游需求强劲。 对于国内，大厂今年加大AI超级入口投入或加快应用普及，需求端的提价空间逐渐增大。 【腾讯控股】25年腾讯云扭亏为盈，差异化竞 ...

Core Insights - The article presents the 2025 China Internet Value List, highlighting the leading applications in various sectors based on user scale and growth rates, particularly focusing on the AIGC (Artificial Intelligence Generated Content) sector and its rapid expansion [2][3]. User Scale Rankings - The top applications by average monthly active users (MAU) from October 2024 to September 2025 include: - WeChat: 109,455.43 million - Taobao: 96,273.09 million - Alipay: 92,643.15 million - Douyin: 88,686.16 million - Gaode Map: 88,685.78 million [5][6]. User Growth Rankings - The top applications with significant user growth (MAU over 50 million) include: - DeepSeek (AIGC): 14,455.28 million, newly launched - 080 (AIGC): 17,204.20 million, with a growth rate of 309.86% - Delta Action (Shooting Game): 5,114.65 million, with a growth rate of 188.57% [9]. Smart TV User Scale - The leading smart TV brands by active device count as of September 2025 are: - Xiaomi: 5,953.76 million - Hisense: 4,615.66 million - Skyworth: 4,423.28 million [11]. OTT Application User Scale - The top OTT applications by active device count include: - Galaxy Qiyi: 12,831.20 million - CIBN Cool Media: 12,483.30 million - Mango TV: 6,548.21 million [12]. AI Application Rankings - The top AI applications by user scale in September 2025 include: - Wenxin Assistant (Baidu AI): 34,729.5 million - 080 (AIGC): 17,204.2 million - Ant Financial Assistant: 8,962.9 million [14]. Advertising Platforms - The leading advertising platforms by user count are: - Tencent Ads: 126,889.06 million - Alibaba Mama: 119,293.23 million - Giant Star Engine: 108,723.63 million [33][34]. Consumer Brand Online Private Domain Traffic - The top consumer brands by online private domain user scale include: - Huawei: 40,563.44 million - Oppo: 26,934.89 million - VIVO: 23,626.35 million [21].

全年预计实现归母净利润2，500万-3，700万元，同比增长128.69%-142.46%，实现扭亏为盈。 业绩拐点核心驱动力来自"锚定大客户"战略的巨大成功，字节跳动与阿里巴巴两大巨头订单引爆增长。 核心驱动力 大客户战略：字节、阿里2025年订单额均大幅增长（2024年已分别拿下2.7亿元、近1亿元）。 依米康业绩预告点评 全年预计实现归母净利润2，500万-3，700万元，同比增长128.69%-142.46%，实现扭亏为盈。 依米康业绩预告点评 大客户战略：字节、阿里2025年订单额均大幅增长（2024年已分别拿下2.7亿元、近1亿元）。 海外主攻：2025年前三季度海外订单近8，000万元，超2024全年4，000多万元，已切入微软、Oracle供应链；马来 西亚分公司已落地。 液冷突破：中标广州天河智算中心一期（约9，000万元），35台机柜采用公司液冷方案，全栈能力获验证。 业绩拐点核心驱动力来自"锚定大客户"战略的巨大成功，字节跳动与阿里巴巴两大巨头订单引爆增长。 国资赋能：引入四川发展战投，2026年资源协同可期。 核心驱动力 ...

【TF计算机】阿里千问3引爆算力需求，IDC产业链再迎强催化 #阿里正式发布万亿参数旗舰推理模型Qwen3-Max-Thinking，总参数量突破1T、预训练数据量达36T Tokens，成为阿 里史上规模最大、能力最强的千问模型。 模型参数与训练数据的极致扩容，直接推动算力需求爆发式增长，为IDC产业链带来确定性增量。 其原生Agent能力的大幅增强，也将加速大模型在各场景落地，进一 【TF计算机】阿里千问3引爆算力需求，IDC产业链再迎强催化 #阿里正式发布万亿参数旗舰推理模型Qwen3-Max-Thinking，总参数量突破1T、预训练数据量达36T Tokens，成为阿 里史上规模最大、能力最强的千问模型。 其原生Agent能力的大幅增强，也将加速大模型在各场景落地，进一步拉动算力产业链景气度上行。 ➡ 万亿参数大模型的训练与推理，对数据中心的单机柜功率、液冷技术及算力调度能力提出更高要求，IDC基础 设施升级迫在眉睫。 阿里算力集群的持续扩张，与"以租代买"模式的深化，让绑定其生态的IDC厂商直接受益。 1 核心IDC供应商：#数据港（阿里系收入占比超80%）、润泽科技、杭钢股份（联合阿里投建浙江云 ...