阿里Qwen 技术负责人 林俊旸的一句"bye my beloved qwen"，在社交媒体上引发了远超一 场普通离职应有的震荡。外界衍生出技术路线分歧、商业化考核高压等种种猜测。 然而，就在3月5日，阿里巴巴CEO吴泳铭的一封内部邮件， 以及来自内部的声音，让事实变 得完整清晰。 2026年3月的第一周，中国AI圈被一则人事变动 刷屏 。 过去两年， 全球 大模型的战事极其惨烈。 伴随而来的是剧烈地战略摇摆和人事变动。字节跳动的AI一号位从朱文佳过渡到吴永辉， OpenAI的11人创始团队仅剩两人留守。 然而， 在与 其他 科技巨头的对比中，阿里Qwen团队此前是业内罕见的"稳定孤岛"，两年未 有重大变动。 这种稳定性让Qwen得以在全球开源社区开疆拓土，衍生模型数突破20万，成为 全球开发者心中的"白月光"。 网络上流传的诸多版本，大多数经不起推敲。所谓阿里放弃开源，子虚乌有，千问的开源策略 从未动摇。所谓基模团队被日活、商业化指标压身，同样查无此事。捕风捉影的消息之所以传 得快，不过是因为它足够好看，足够像一出戏。 但真实发生的事，远没有那么戏剧化。千问从一个技术项目跃升为阿里的集团核心战略，量级 变了 ...

大事观察： 所罗门指出，市场情绪部分得益于美国对确保霍尔木兹海峡航运安全的承诺。但该海峡实际已被伊朗海军控制，十多艘油轮被击中，航 运风险极高。 冲突影响正蔓延至全球。日韩股市遭受严重冲击，韩国KOSPI指数暴跌12%。摩根大通警告，若海峡持续关闭，伊拉克和科威特原油供 应可能数日内大幅减少。 2025年成国产人形机器人关键年 十四届全国人大四次会议今天（3月4日）在人民大会堂举行新闻发布会，大会发言人娄勤俭就大会议程和人大工作相关问题回答中外记者提问。 娄勤俭说："2025年是国产人形机器人产业实现技术突破与场景落地双重跨越的关键一年。人形机器人技术能力、应用场景的不断拓展，让人们看到一个 充满无限可能的未来。" 中国 智能眼镜 销量暴涨211% 高盛CEO所罗门：市场对伊朗战争反应"温和"令人意外 高盛集团董事长兼CEO大卫·所罗门在悉尼峰会上表示，金融市场对中东冲突的反应"较为温和"，这令他感到意外。他认为要全面了解局 势并消化影响，可能还需数周时间。 冲突已引发油价飙升与通胀担忧。布伦特原油价格突破每桶82美元，创2020年以来最大两日涨幅。交易员因此大幅下调对美联储降息的 预期。 根据洛图科技（RU ...

Core Insights - The article discusses the evolution of AI models, particularly focusing on the rise of reasoning models and their impact on decision-making processes, highlighting a shift from OpenAI's dominance to emerging Chinese models [1][3][25]. Group 1: Reasoning Models - OpenAI initiated a "reasoning revolution" in September 2024 with the launch of models like o1 and o1-mini, which have since become a standard feature across major AI labs [3]. - By 2025, every notable AI lab released at least one reasoning model, with some offering hybrid models that can switch between reasoning and non-reasoning modes [4][5]. - The true value of reasoning models lies in their ability to drive tools, enabling multi-step task planning and execution, significantly improving AI-assisted search capabilities [5][6]. Group 2: Programming Agents - 2025 is characterized as the year of programming agents, with the release of Claude Code marking a significant advancement in this area [11][12]. - Programming agents can write, execute, and debug code, demonstrating exceptional performance in identifying bugs within complex codebases [7][10]. - The CLI programming agent model gained traction, with various labs launching their own versions, indicating a growing interest in command-line access to AI models [13][17]. Group 3: Subscription Models - The emergence of subscription plans, such as Claude Pro Max at $200 per month and OpenAI's ChatGPT Pro, has generated substantial revenue, although specific user data remains undisclosed [23][24]. - Users have expressed willingness to pay higher subscription fees for advanced capabilities, particularly when engaging in more complex tasks that consume tokens rapidly [24]. Group 4: Chinese AI Models - In 2025, Chinese AI labs made significant strides, with models like GLM-4.7 and DeepSeek gaining prominence, leading to a shift in the global AI landscape [25][28]. - The release of DeepSeek 3 in late 2024 triggered a market reaction, causing a significant drop in NVIDIA's market value, highlighting the impact of Chinese models on investor sentiment [28]. Group 5: Long Tasks and Image Editing - AI models have shown remarkable progress in handling long-duration tasks, with capabilities doubling approximately every seven months, as evidenced by the performance of models like GPT-5 and Claude Opus 4.5 [31][33]. - The introduction of prompt-driven image editing features in ChatGPT led to a rapid increase in user adoption, showcasing the potential for consumer-level applications [34][35]. Group 6: Competitive Landscape - OpenAI's position as a leader in the LLM space is being challenged by competitors like Google Gemini, which has released multiple iterations of its models with competitive pricing and capabilities [46][47]. - The competition is intensifying, particularly in image generation and programming capabilities, with Google leveraging its proprietary TPU hardware to enhance model performance [47][48].

Core Insights - The article discusses how Chinese open-source AI models are gaining popularity among American companies, suggesting that China may be quietly winning the AI race [1] Group 1: Adoption of Chinese AI Models - Pinterest is utilizing Chinese AI models, specifically the DeepSeek R-1 model, to enhance its recommendation engine, indicating a growing trend among U.S. companies to adopt these technologies [2] - The CEO of Pinterest, Bill Ready, noted that the open-source nature of DeepSeek has sparked a wave of interest in open-source AI models [2] - Other Chinese open-source models mentioned include Alibaba's Qwen and Moonshot's Kimi, with ByteDance also developing similar technologies [2] Group 2: Advantages of Chinese AI Models - The CTO of Pinterest, Matt Madridigal, stated that the ability to download and customize these models provides a significant advantage over models from companies like OpenAI, which are not freely available [2] - The accuracy of models trained using open-source technology is reported to be 30% higher than that of off-the-shelf models, with lower optimization costs [2] Group 3: Broader Industry Impact - Chinese AI models are recognized by numerous Fortune 500 companies, with Airbnb's CEO Brian Chesky highlighting the benefits of Alibaba's Qwen for their AI customer service agents, citing quality, speed, and cost-effectiveness [4] - The Hugging Face platform shows that Chinese models occupy multiple spots in the top ten most popular models, with Alibaba's Qwen surpassing Meta's Llama as the most downloaded large language model [4] Group 4: Competitive Landscape - A report from Stanford University indicates that Chinese AI models have either caught up to or surpassed global counterparts in terms of capability and user base [6] - The success of China's open-source model development is partially attributed to government support [6]

Core Insights - The article highlights a significant shift in the global AI industry, indicating that 2026 will be a pivotal year for Alibaba's Qwen model, as it gains traction against competitors like OpenAI's GPT-5 and Google's Gemini 3 [1][3][6] Industry Trends - The performance of Chinese AI models such as Qwen, DeepSeek, and others is increasingly recognized, with their flexibility and developer-friendly nature contributing to their rising popularity [3][6] - The article notes a transition in Silicon Valley, where companies are beginning to favor Chinese models for their cost-effectiveness and performance, as exemplified by Airbnb's CEO praising Qwen over OpenAI's offerings [6][7] Technological Advancements - Chinese AI models are not only competing on price but are also advancing in technology, with a commitment to openness and continuous improvement, contrasting with the more closed-off approach of some American companies [7][9] - The rise of Chinese models reflects a new industry standard that prioritizes application breadth and deployment flexibility over mere parameter size and conversational intelligence [9][12] Capital Market Dynamics - MiniMax and Zhizhu AI have successfully attracted significant international investment, validating the global value of Chinese AI as a core asset, with MiniMax securing approximately $350 million from 14 cornerstone investors [13][14] - The article emphasizes that the recognition of Qwen by a leading tech media outlet signifies a broader acceptance of Chinese AI models in the global market, suggesting a collective movement towards international expansion [14]

阿里千问（Qwen）是由阿里巴巴集团推出的开源大语言模型系列，自2023年8月起，陆续开源四代模 型。豆包则是字节跳动公司推出的AI工具，同样上线于2023年8月。 12月19日，阿里千问在其公众号发文称，自己"被AI整了"，此前在网络流传的 " 千问全员会，吃豆包 " 一图，系AI生成。不仅广场大会是假的，画面中的Logo与工牌也全是错的。 END 点击关键字可查看 潜望系列深度报道丨 股事会专栏 丨 投资小红书 丨 e公司调查 丨 时报会客厅 丨 十大明星私募访谈 丨 网信办和证监会联手，一批涉资本市场不实信息的账号被处置 丨 突发！000592、603122，尾盘 直线跳水！ 丨 利好！涨停潮！ 丨 昨夜，这一赛道，大爆发！美联储，降息大消息！ 丨 重磅数据发 布！美股拉升！ 丨 吴清最新发声 丨 603288，拟每10股派3元！ 丨 稀土，大消息！商务部回应！ 丨 这一板块，爆发！多股涨停 丨 昨夜，美股大跌，科技股跳水！美联储，最新发声！ 丨 突发！阿 里创投减持 责编：叶舒筠 校对： 姚远 版权声明 证券时报各平台所有原创内容，未经书面授权，任何单位及个人不得转载。我社保留追 究相关 行 为主体 ...

Group 1 - Meta is training a new model codenamed "Avocado" using Alibaba's Qwen model for distillation optimization, indicating a strategic shift in their AI development approach [1][3] - The team at Meta is utilizing multiple third-party models, including Google's Gemma and OpenAI's gpt-oss, as part of the training process for "Avocado" [3] - Mark Zuckerberg's previous concerns about the potential censorship of Chinese models have shifted, as he now incorporates Chinese technology into Meta's AI strategy, reflecting a significant change in attitude [3] Group 2 - Alibaba's Qwen model has gained recognition as a leading global open-source model, becoming a reference point for major tech companies like Meta in their pursuit of industry leadership [3] - Since its public testing began on November 17, the Qwen App has achieved over 30 million monthly active users within just 23 days, marking it as one of the fastest-growing AI applications globally [4]