Core Insights - Alibaba has quietly launched two new models, Qwen3.5-Plus and Qwen3.5-397B-A17B, on the chat.qwen.ai platform [1] - Qwen3.5-Plus is positioned as the latest large language model in the Qwen3.5 series, while Qwen3.5-397B-A17B is the flagship model of the open-source Qwen3.5 series [1] - Both models support text and multimodal tasks [1]

当人们平心静气深入思考过去、当下和未来，其实不难发现，所谓稳定的秩序和格局，从来都充满张力暗流涌动，阶段性扰动乃至重组是常态，"帕累托 改进"你好我好全都好的繁华年代则不常见。 编者按：随着2025年的结束，这个世纪已过去四分之一个年头。这四分之一个世纪给我们留下了什么?或许目前为止，更多的是变化中的措手不及和对未 来的茫然。曾经脚下坚实的地面正在一点点偏移、开裂，所有人都亟需找到新的立足点。 好在还有阅读。任时代如何变化，书籍永远可以给予我们启示，甚至可以说，书籍正是审视变动的产物。一代代写作者都在尝试给出自己的回答。而我 们，作为阅读者，也同时承受着这份考验。 《欧洲告急》把丘吉尔和奥威尔在二战前后的命运交叉展现，乍看之下颇令人不解，二人的政治和职业生涯交叉之处并不算多，顶多算是平行或勉强的前 后脚(比如在印度和缅甸的殖民地时光)关系。政治立场也迹近光谱的两端，一个偏左，一个偏右。 经观书评推出新年特辑"寻找新的立足点"，邀请各界读书人分享他们在2025年的阅读与思考。本期作者是作家邓安庆，他说： "我相信，时间的确有如此淘洗的伟力。当时间之河毫不留情地流向2026年时，每一个普通人都将继续努力过好自己的 ...

Core Viewpoint - The Chongqing Library has opened an "Anti-Japanese War Literature Smart Experience Zone," integrating digital technology with historical documents to enhance public engagement and understanding of wartime history [1][5]. Group 1: Experience Zone Features - The experience zone features a "digital librarian" powered by a large language model, providing real-time explanations and knowledge services related to anti-war literature [2]. - A "Cultural Corridor" presents a timeline of significant events from the Anti-Japanese War, allowing readers to interact with historical materials [2]. - Virtual reality (VR) technology enables readers to immerse themselves in historical settings, such as the Roosevelt Library, and experience period costumes for an interactive learning experience [2]. Group 2: Technological Integration and Impact - The Chongqing Library has transitioned from static preservation of literature to dynamic service, utilizing over 30,000 anti-war documents for digital organization and intelligent application [5]. - The integration of smart technologies enhances the reading experience and transforms the way public libraries organize and disseminate knowledge [5]. - The establishment of the experience zone reflects a commitment to making unique collections accessible to the public, thereby increasing cultural awareness and participation [5]. Group 3: Future Developments - The Chongqing Library plans to continue advancing its smart library initiatives, improving its specialized literature knowledge base, and expanding intelligent service scenarios [5].

2026-02-12 摘要 智谱 GLM 5 在长期任务中取得显著进步，缩小了与 Cloud OPIUS 4.5 的差距，并在部分 benchmark 测试中超越后者。GLM 5 的 API 和 Coding Plan 已上线，API 定价已公布，Coding Plan 将于 2 月 16 日 正式上线。 智谱推出了特工模式（视角转换、数据分析、AI PPT 支持）、数据洞察 （数据解析、画图、导出多种格式文件）和写作功能（支持多种写入模 式并导出 PDF 和 Word）。 Minimax 发布了 M2.5，是首个为 Agent 场景原生设计的生产级模型， 支持 Excel 处理、深度调研及 PPT 等生产力场景，但 API 尚未上线， 也未公布价格。 GLM 5 的 Output 定价为每百万 Tokens $3.2，GLM 4.6 和 GLM 4.7 从 $1.5 提升至 $2.2，国内 API 价格提升幅度在 33%至 100%之 间，海外市场为 67%至 100%。Coding Plan 国内 Light 和 Max 两个 季度收费版本分别提高了 23%和 18%；海外市场分别提高了 60%和 30% ...

中金研究 大模型的演进史，本质上是一部与"遗忘"抗争的历史。 当我们惊叹于模型的推理能力时，往往忽视了一个重要短板： 在缺乏记忆留存的架构下，模型 每一次对历史信息的处理，本质上都是一次昂贵的"重复计算"。 这种以高昂算力对抗遗忘的粗放模式，正面临着显存墙与上下文窗口的物理极限。我 们认为，2026年及之后的AI Infra主战场将增加"模型记忆"这一极。 何为模型记忆？如何理解短期、中期、长期记忆三层记忆系统对应的软硬件需求？ 如何对应模型训练、推理、Agent场景理解记忆分层系统？我们将在本报告中予以解答。 点击小程序查看报告原文 Abstract 摘要 短期记忆构成大模 型单 次推理的"当前视野"。 作为高频读写、对延迟极度敏感的"热数据"，其核心矛盾在于KV Cache对显存容量与带宽的双重挤占。软 件端通过PagedAttention显存虚拟化与PD分离调度进行优化，并探索出无限注意力（Infini-attention）等前沿架构以支撑百万Tokens上下文窗口。这一逻辑 直接锚定了HBM与片上SRAM作为突破"显存墙"与"延迟墙"的重要硬件要素。 中 期记忆保障跨会话的情景连续性，是Agent的基 ...

Core Insights - The article discusses the delivery performance of Li Auto in January 2026, highlighting a total delivery of 27,668 vehicles, with 9,358 being range-extended and 18,310 being pure electric [1][2] - It mentions a significant organizational restructuring within Li Auto, transitioning from a Huawei IPD model to a Toyota CE model [1] - The article also notes the introduction of two product lines within Li Auto, with specific leadership assigned to each line [3] Delivery Performance - In January 2026, Li Auto delivered a total of 27,668 vehicles, which includes 9,358 range-extended vehicles and 18,310 pure electric vehicles [1][2] - The delivery numbers for previous months show fluctuations, with December 2025 having the highest at 44,246 vehicles, and a notable decrease in January 2026 [2] Organizational Changes - Li Auto is restructuring its product lines into two main categories: one focusing on MEGA, L9, L8, and L7, and the other on the i series and L6 [3] - Leadership changes include汤靖 overseeing the first product line and 李昕旸 managing the second [3] Market Position and Strategy - The article indicates that Li Auto is focusing on enhancing user experience over the initial purchase experience, as stated by the company's founder [3] - There is a mention of Li Auto's strategic goal for 2028, emphasizing the importance of AI and agent technology in their future plans [4] Industry Context - The article references a broader industry context where Li Auto is seen as a leader in self-developed materials and applications within the automotive sector [4] - It also highlights the competitive landscape, noting that the performance of other brands in the market, such as AITO, has been better in the 300,000+ market segment compared to Li Auto [3][5]

近日，瑞银发布博通(AVGO.US)专项研报，维持其"买入"评级及475美元目标价不变。报告围绕TPU(张量处理单元)需求激增展开，指出LLM开发者加 速推进定制ASIC路线，TPU作为GPU的中间替代方案需求显著增长，成为博通业绩增长的核心驱动力。 除谷歌外，Anthropic、META等TPU核心客户与传统云服务客户存在显著差异：这类企业可完全掌控自身的软件栈，因此对英伟达统一计算架 构(CUDA)的依赖度远低于传统企业级云服务客户。 经过一系列供应链调研分析，该行优化了自下而上的定制化专用集成电路模型，目前预测博通公司2027年将出货超500万颗张量处理单元 (2026年出货量约为370万颗);在2028年v8ax(太阳鱼)型号成为出货主力前，2027年出货的产品中略超半数为v7(铁木)型号。上述两款产品均 基于台积电3纳米工艺打造，该行认为凭借台积电充足的晶圆供应配额，博通能够充分把握这一需求增长机遇。 该行当前预测，博通2026财年人工智能业务营收约为600亿美元(同比增长约200%)，2027财年将增至约1060亿美元(同比增长约80%)， 2028财年进一步升至约1500亿美元。定制化计算业务营 ...

Core Insights - DeepSeek is advancing its new model version with a grayscale test, potentially the final version before the official V4 launch [1] - The V4 model is expected to be released in mid-February 2026, and it will not replicate the global AI computing demand panic seen during the V3 launch [2] - The core value of V4 lies in driving the commercialization of AI applications through underlying architectural innovations rather than disrupting the existing AI value chain [2] Model Enhancements - The context length of the model has been expanded from 128K to 1M, nearly a tenfold increase, and the knowledge base has been updated to May 2025 [1] - V4 is expected to introduce two innovative technologies, mHC and Engram, which aim to overcome computing chip and memory bottlenecks [2][8] - Initial internal tests indicate that V4 outperforms models like Anthropic Claude and OpenAI's GPT series in programming tasks [2] Technical Innovations - mHC (Manifold Constraint Hyperconnection) addresses the bottlenecks in information flow and training instability in deep Transformer models, enhancing the richness and flexibility of communication between neural network layers [4] - Engram is a "conditional memory" module that decouples memory from computation, allowing static knowledge to be stored in a sparse memory table, thus freeing up expensive GPU memory for dynamic calculations [6] Cost Efficiency and Market Impact - The introduction of mHC and Engram is expected to significantly reduce training and inference costs, stimulating downstream application demand and initiating a new cycle of AI infrastructure development [8] - The report suggests that Chinese AI hardware manufacturers may benefit from increased demand and investment due to these cost optimizations [8] Market Dynamics - The market landscape has shifted from a dominant player to a more fragmented competition, with DeepSeek's market share declining as more players enter the field [9][11] - The efficiency in computing management and performance improvements from DeepSeek are accelerating the development of Chinese large language models and applications, altering the global competitive landscape [11] Opportunities for Software Companies - Major global cloud service providers are actively pursuing general artificial intelligence, and the capital expenditure race continues [12] - If V4 can maintain high performance while significantly lowering training and inference costs, it will help developers convert technology into revenue more quickly, alleviating profit pressures [12] - Enhanced capabilities of V4 are expected to create more powerful AI agents, transforming them from mere conversational tools to capable assistants that can handle complex tasks [12]

多名用户反馈，DeepSeek在网页端和APP端进行了版本更新，支持最高1M（百万）Token的上下文长度。 而去年8月发布的DeepSeekV3.1上下文长 度拓展至128K。 《科创板日报》记者实测中发现，DeepSeek在问答中称自身支持上下文1M，可以一次性处理超长文本。记者在提交了超过24万个token的《简爱》小 说文档，DeepSeek可以支持识别文档内容。 ...

Core Viewpoint - The article discusses the development of AMIE, an AI system designed to assist in complex cardiology care, addressing the shortage of specialized medical professionals and improving diagnostic accuracy and efficiency [3][7][16]. Group 1: Challenges in Cardiology - There is a significant shortage of specialized medical knowledge in cardiology, leading to challenges in providing timely and effective medical services [2]. - The World Health Organization (WHO) predicts a global shortage of 18 million healthcare workers by 2030, with cardiology being particularly affected [5]. - In the U.S., over half of the states lack specialized centers for hypertrophic cardiomyopathy, resulting in 60% of patients not receiving a diagnosis [5]. Group 2: AMIE Development and Functionality - AMIE (Articulate Medical Intelligence Explorer) is an experimental AI system based on the Gemini 2.0 Flash large language model, specifically designed for complex cardiology cases [7][8]. - Unlike traditional AI systems, AMIE can analyze multiple diagnostic tests, including ECG, echocardiograms, and cardiac MRIs, providing comprehensive diagnostic suggestions [8]. Group 3: Clinical Trial Design and Results - The study utilized a randomized controlled trial design, selecting clinical data from 107 real patients with various complex heart conditions [10]. - Nine cardiologists were divided into two groups: one using AMIE for diagnostic assistance and the other relying solely on personal experience [11]. - Results showed that cardiologists preferred the diagnoses assisted by AMIE, with a preference rate of 46.7% compared to 32.7% for those relying on personal experience [15]. Group 4: Impact on Diagnostic Quality and Efficiency - AMIE significantly reduced clinical errors, with a notable decrease in significant errors (13.1% vs 24.3%) and important omissions (17.8% vs 37.4%) [15]. - The use of AMIE improved clinical assessments in 57% of cases and saved time in 50.5% of cases, enhancing doctors' confidence in their decisions [15]. Group 5: Future Implications of AMIE - AMIE excels in management plan formulation and diagnostic test recommendations, providing detailed information on rare diseases and prompting critical thinking [16]. - The study highlights the potential for AI to enhance human expertise in specialized medical fields, particularly in areas with a shortage of specialists [16]. - This research marks a significant step for AI in specialized healthcare, indicating a future direction of human-AI collaboration for improved patient care [16].