不圆 发自 凹非寺 量子位 | 公众号 QbitAI 百度最新视频生成模型 蒸汽机2.0 （MuseSteamer 2.0），好像真的有点东西。 这是在网上热传的一段由它生成的视频，可以说是要声音有声音，要画面有画面，不说的话还以为是某部重生剧的先导片。 AI配音的中文非常自然，和角色口型也对得很好。 我们也试着生成了一个小视频，仅用1张图片和1段提示词，就做出了这样的效果： 仔细听，这只猫甚至会呼噜噜，远处还有虫子叫。 网友评价：这简直像魔法一样！ 它要怎么用才会更好玩？又能用来做什么呢？ 我们实测了这款模型，一起来看它的具体表现。 模型表现 该说不说，作为全球首个 中文 音视频一体化生成的I2V模型，蒸汽机模型在中文语音的表现上可以说是手拿把攥，但这是蒸汽机1.0模型刚出 的时候就已经介绍的东西。 作为升级版本，蒸汽机2.0更加擅长 复杂运镜 ，用镜头讲故事的能力也更强，画质进一步提升。 让我们看看，作为普通人能用这个模型实现什么想法？ 它的表现 和爆火的Veo3相比 ，哪个更好呢？ 画画人狂喜：绘画转视频 我们让豆包生成了一张手绘风格的图片，画面上是一只大野兔蹲在草丛里。 就假装它是我们画出来的吧 （手 ...

2025 年 08 月 25 日 DeepSeek-V3.1 开启 AI 高效计算时代，百度发 布音视频一体化模型蒸汽机 2.0 分析师：宝幼琛 S1050521110002 baoyc@cfsc.com.cn 行业相对表现 | 表现 | 1M | 3M | 12M | | --- | --- | --- | --- | | 计算机(申万) | 14.6 | 30.8 | 101.2 | | 沪深 300 | 8.3 | 15.1 | 34.3 | 8 月 21 日，美国机器人技术企业 FieldAI 成功完成 4.05 亿美 元融资，由英伟达旗下 NVentures 和贝索斯家族办公室 Bezos Expeditions 共同领投，Khosla Ventures、淡马锡及英特尔 资本等机构跟投，投后估值达 20 亿美元。该公司专注于物理 AI 与机器人自主技术，其核心"Field 基础模型"（FFMs） 以物理优先为设计原则，专用于复杂和非结构化环境中的机 器人智能决策，支持多种硬件形态且无需依赖预建地图或 GPS。目前技术已应用于全球数百个工业场景，涵盖建筑、能 源、制造和物流等领域，展现出实体智能领域的 ...

AI视频生成正成为大模型竞争的核心战场，成本与质量的平衡逐渐成为行业比拼的关键。 8月21日，百度在"热AI大会"上发布蒸汽机2.0，Turbo、Lite、Pro及有声版同步上线。新版本在语音与 画面同步、多角色对话生成、中文场景适配等方面进行了升级，并将定价大幅下调，刊例价相比同类产 品下降至70%，让好莱坞百万级特效成本降至"百元"。 据百度副总裁、移动生态商业体系负责人陈一凡透露，降本背后是百度长期在GPU算力架构和工程优化 上的积累。"自2016年起，商业研发团队就率先在搜索广告场景中引入GPU，形成了软硬件结合的技术 路径。本次蒸汽机迭代，依托百度智能云'百舸'平台与自研昆仑芯片，与策略工程架构和底层算力结 合，推理效率和算力利用率得到大幅提升，从而支撑了价格下探。" 在国内外厂商加速迭代视频生成应用的背景下，百度蒸汽机选择以"技术突破+价格下行"同时发力，意 在撬动更大规模的创作者和商业市场。 技术突破与产品升级：从"一体化"到"可用性" 相比文字和图像生成，视频生成的难点在于多模态的统一：画面要连续自然，声音要真实可信，更重要 的是口型、表情、动作与语音节奏要完全对得上。百度商业研发首席架构师李 ...

以下文章来源于快鲤鱼 ，作者巴里 快鲤鱼 . 创业邦旗下AGI矩阵号，寻找海内外创新性的AGI高成长公司，记录AGI商业领袖的成长轨迹。 全球AI产业周报 为你精选过去一周（8.16-8.22）最值得关注的AI新闻和 国内外热门AI投融资事件 ，帮助大家及时了解全球AI市场动向。 本周AI热点资讯 国内大事 DeepSeek V3.1发布：更强的Agent能力，更贵的API 8月21日，DeepSeek 正式发布新模型V3.1，被官方称为「迈向智能体时代的第一步」。虽然未见期待已久的R2模型，但本次迭代重点在于更强的Agent 能力、混合思考模式与更高思考效率。 V3.1采用混合推理架构，用户可在「思考模式」与「非思考模式」间自由切换：复杂任务可调用深度推理，简单任务则快速响应，避免性能浪费。官方测 试显示，V3.1-Think在输出token数减少20%-50%的情况下，表现与此前R1-0528持平甚至更快。 此外，新模型在工具调用和智能体任务中的表现明显提升，编程与搜索Agent测评均优于前代。基础模型在V3的基础上新增8400亿tokens训练，并已在 Huggingface与魔搭开源。DeepSeek ...

Key Points - The article discusses various developments in the tech and automotive industries, highlighting significant corporate actions, product launches, and market strategies. Group 1: Corporate Developments - Former CFO of Zhibo Network publicly criticized the company's upcoming IPO, stating that he left due to a lack of confidence in the business and accused certain executives of being opportunistic [4][6]. - Alibaba announced the spin-off of Zhibo Network for an independent listing on the Hong Kong Stock Exchange, with plans to retain over 30% ownership post-IPO [6]. - Alibaba's Lingxi Entertainment has shifted its reporting structure to report directly to CFO Xu Hong, indicating potential changes in business strategy [12][13]. Group 2: Product Launches and Innovations - NIO unveiled the new ES8 model, with a starting pre-sale price of 416,800 yuan, featuring significant upgrades in size and technology [19]. - Vivo introduced the Vision Exploration Edition, the lightest MR headset in the industry, weighing only 398g, designed for enhanced user experience [30]. - DeepSeek released version 3.1, which includes significant upgrades and price adjustments for its API services, reflecting a shift towards next-generation domestic chips [11]. Group 3: Market Strategies - Alibaba's local services division is launching a new group-buying feature called "Flash Group," aimed at price-sensitive consumers, to compete with Meituan's similar offerings [18]. - Multiple ride-hailing platforms, including Didi and T3, have announced reductions in commission rates to support driver income and expand platform capacity [24][25]. - Zero Run Auto reported a cumulative delivery of over 900,000 vehicles, achieving profitability in the first half of the year and adjusting its annual sales target upwards [26][27]. Group 4: Financial Performance - Kuaishou reported a revenue of 35.05 billion yuan for Q2 2025, with a net profit increase of 20.1%, and announced a special dividend for shareholders [39]. - Bilibili's Q2 revenue reached 7.34 billion yuan, with significant growth in advertising and gaming revenue, and a record high in user engagement metrics [40]. Group 5: Competitive Landscape - Samsung's HBM4 samples have passed initial testing with Nvidia and are set to enter pre-production, potentially challenging SK Hynix's dominance in the AI memory chip market [44][45]. - Intel is negotiating with large investors to replicate a previous financing deal with SoftBank, aiming to bolster its capital structure [46]. Group 6: Privacy and Regulatory Issues - Meta is facing allegations of circumventing Apple's privacy restrictions to enhance ad revenue, with claims of misleading advertisers about the performance of its Shop Ads [51][52]. - xAI's Grok platform experienced a significant privacy breach, exposing over 370,000 user chat records due to design flaws in its sharing functionality [46][47].

Group 1 - DeepSeek officially released DeepSeek-V3.1 on August 21, featuring a hybrid reasoning architecture, improved thinking efficiency, and enhanced agent capabilities. The new model supports both thinking and non-thinking modes, providing faster responses compared to DeepSeek-R1-0528 [1] - The official app and web model have been upgraded to DeepSeek-V3.1, allowing users to switch freely between thinking and non-thinking modes via a "deep thinking" button [1] - DeepSeek announced a price adjustment for API calls starting from September 6, 2025, and will eliminate night-time discounts. All API services will continue to be billed at the current rates until that date [3] Group 2 - Tesla launched a new six-seat Model Y in China, priced at approximately $47,200, with CEO Elon Musk indicating that this variant may not be produced in the U.S. due to the rise of autonomous vehicles [5] - Kuaishou reported a 13.1% year-on-year increase in total revenue for Q2 2025, reaching RMB 35 billion, with adjusted net profit growing by 20.1% to RMB 5.6 billion [11] - Xiaopeng Motors' chairman He Xiaopeng purchased 3.1 million shares at an average price of HKD 80.49, increasing his total ownership to approximately 18.9% [11] - Sohu's CEO Zhang Chaoyang stated that Sohu Video will not participate in short drama production, focusing instead on long dramas and live broadcasts [11] - NIO announced the pre-sale of its new ES8 model starting at RMB 416,800, with deliveries expected to begin in late September 2025 [23] Group 3 - Meta responded to rumors of freezing AI department hiring, clarifying that it is a basic organizational adjustment while establishing a framework for new AI projects [9] - KKR is reportedly the leading bidder for Nissan's global headquarters building, offering approximately $610 million [16] - Intel is negotiating with large investors to raise capital through discounted equity offerings [16] - Nuro completed a $203 million Series E funding round, achieving a valuation of $6 billion [18]