开源大模型

Search documents
阿里云领投硅基流动A轮融资 半年融资两轮背后:开源大模型崛起带来业务爆发式增长
Mei Ri Jing Ji Xin Wen· 2025-06-09 12:35
Core Insights - SiliconFlow, an AI startup, has completed a Series A financing round of several hundred million RMB, led by Alibaba Cloud, with participation from existing investors like Innovation Works and financial advisory from Huaxing Capital [1] - The company has experienced explosive growth in business this year due to the rise of open-source large models like Alibaba's Tongyi Qwen and DeepSeek, alongside a surge in demand for AI inference computing power [1] - The funding will be used to increase R&D investment and expand both domestic and international markets [1] Company Overview - SiliconFlow aims to provide developers with essential tools for application innovation based on AI models, promoting "token freedom" for developers [3] - The company has launched its SiliconCloud platform, which features a full version of the DeepSeek R1/V3 model, successfully deploying it on domestic chips [3] - SiliconFlow's user base has surpassed 6 million, with thousands of enterprise clients generating over a trillion tokens daily [3] Product and Service Offerings - The company offers a range of solutions including API services, dedicated instances, software subscriptions, and integrated large model machines, serving major clients across various industries such as internet, finance, manufacturing, and entertainment [4] - SiliconFlow is focused on reducing the barriers for developers and enterprises in AI application development and deployment [4] Market Potential - The large model sector presents significant market opportunities, particularly in B2B services, as many enterprises are leveraging large models for specialized services [4] - There is a strong demand for fine-tuning and inference of large models, which SiliconFlow is well-positioned to capitalize on [4]
最早接住DeepSeek流量的硅基流动,新获阿里领投数亿元融资|36氪独家
36氪· 2025-06-09 10:47
以下文章来源于暗涌Waves ,作者暗涌 暗涌Waves . 钱的流向,人的沉浮。36氪旗下投资报道账号。 "一系列偶然又必然的选择。" 文 | 于丽丽 来源| 暗涌Waves(ID: waves36kr ) 封面来源 | IC Photo " 暗涌Waves " 获悉,AI Infra公司硅基流动新近完成一轮由阿里云领投的数亿元人民币融资。老股东创新工场等机构超额跟投,华兴资本担任独家财务顾 问。更早入局的,还包括美团(战略投资)、创新工场、耀途资本、奇绩创坛、华创资本、普华资本等机构。 硅基流动创始人袁进辉把本次融资比作"一次双向奔赴"。拆解看,阿里对AI基础设施一直是战略级投入。年初,阿里CEO吴泳铭就宣布了在云和AI硬件 基础设施领域的庞大投资:3800亿。这也是中国民营企业有史以来在此领域的最大规模投资纪录。而从硅基流动角度,除了获得融资外,未来也"可以与 阿里巴巴通义千问有更好的生态协作,还能在算力、国内外市场扩展等方面广泛协作。"袁进辉如此说。 这也是硅基流动在获得爆发式增长后的一次融资。上次融资还是爆发前的2024年底,当时硅基流动完成的是,华创资本领投,普华资本跟投,老股东耀途 资本超额跟投 ...
2025年第18期(总899期):开源大模型DeepSeek实现三个“首
Sou Hu Cai Jing· 2025-06-07 08:35
今天分享的是:2025年第18期(总899期):开源大模型DeepSeek实现三个"首次",应借助开源顺势推动AI普惠化平权化发展 报告共计:10页 开源大模型DeepSeek的创新实践与AI普惠化发展路径 一、DeepSeek:全球开源AI大模型的新标杆 AI大模型开源需满足代码完整、模型参数公开、训练数据透明三大核心标准,较传统软件开源更复杂。此前多数大模型厂商走 纯闭源或"半开源"路线,如OpenAI的GPT-4、Meta的Llama 3仅部分开源且附带商用限制,仅有少数机构实现全栈开源。 DeepSeek则以全栈开源和宽松协议树立新典范:不仅开放代码、权重、文档下载,公开GPRO训练算法等技术细节,还采用无商 用限制的MIT许可,支持用户进行"模型蒸馏",为行业提供了透明、开放的技术基座。 二、DeepSeek的三大突破性"首次" 1. 技术路径革新:开辟大模型发展第二路线 DeepSeek-R1通过纯强化学习(RL)训练证明"小而美"路径的可行性,打破了依赖"Scaling Law"的"唯资源论"定式。其推理成本 与定价显著低于国际主流模型,为资源有限的国家提供了低成本高效能的技术方案,助力缩小全球 ...
明线为AI应用起势,暗线为文化自信,游戏板块反弹上攻趋势显著,聚焦游戏板块布局机会
Mei Ri Jing Ji Xin Wen· 2025-06-03 03:11
Group 1 - The gaming sector is experiencing a strong recovery, with the gaming ETF (159869) rising nearly 4% as of the report, and has seen net inflows in 4 out of the last 5 trading days, indicating sustained investor interest [1] - In May, the National Press and Publication Administration approved 130 domestic and 14 imported online games, totaling 144 approvals, which marks a new monthly record in nearly two years [1] - According to Gamma data, the Chinese gaming market is projected to reach 273.51 billion yuan by April 2025, representing a year-on-year growth of 21.93%, driven by mobile games and overseas revenues [1] Group 2 - Huachuang Securities highlights that IP toys and live performances are key growth areas in the new consumption sector, with expectations for continued rapid growth in the industry [2] - The media sector is seeing a rise in AI applications, with 2023 anticipated to be a pivotal year for the explosion of open-source large models in China [2] - By 2024, global gaming industry revenue is expected to reach 187.7 billion USD, with China accounting for over 30% of this revenue, and self-developed games making up over 80% of the domestic market [2]
传媒行业周观察(20250526-20250530)
Huachuang Securities· 2025-06-03 00:25
Investment Rating - The report maintains a "Recommendation" rating for the media industry, expecting the industry index to rise more than 5% over the next 3-6 months compared to the benchmark index [49]. Core Viewpoints - The report expresses a positive outlook on the IP toy sector, highlighting its long-term growth potential driven by diverse product categories. The recent success of the "Jinli Naju" limited edition merchandise from Alibaba Pictures during the Dragon Boat Festival is noted as a significant indicator of market interest [5][6]. - The media sector is currently experiencing a resurgence in AI applications, with a focus on cultural confidence stemming from popular IPs like "Nezha." The report anticipates a reshaping of the application landscape in 2023, particularly in public cloud services and B-end SaaS enterprises [5][6]. - The gaming market is highlighted as a key area of interest, with recommendations to focus on companies like Huatuo, Perfect World, and JiBit, driven by product cycles and deepening AI integration [5][6]. Summary by Sections Market Performance Review - The media sector index rose by 1.74% last week, outperforming the CSI 300 index, which fell by 1.08%, resulting in a relative outperformance of 2.82% [8]. - The total market capitalization of the media sector is approximately 1,569.05 billion yuan, with 140 listed companies [2]. Gaming Market - Tencent's games dominate the iOS sales rankings, with "Honor of Kings" and "Peacekeeper Elite" leading the charts. New releases from other companies are also noted, indicating a competitive landscape [16][17]. Film Market - As of May 30, 2025, the film market has achieved a box office of 24.545 billion yuan, recovering approximately 98% of the box office compared to the same period in 2019. The total number of viewers is around 588 million, recovering about 86% [19][22]. - The top films during the week of May 26 to May 30 include "Mission: Impossible 8" and "Lilo & Stitch," with significant box office contributions [26]. Key Company Announcements - Meituan reported a revenue of 86.6 billion yuan for Q1 2025, exceeding market expectations by 18.1%, with a net profit of 10.95 billion yuan, reflecting a year-on-year growth of 46.2% [33]. - Kuaishou's Q1 2025 revenue reached 32.608 billion yuan, showing an 8.8% year-on-year increase, with a net profit of 3.978 billion yuan [34].
“开源大模型之城”,为何是杭州?
Sou Hu Cai Jing· 2025-05-30 07:09
在软件领域,开源与闭源两种路线之争由来已久。此前大模型以闭源为主,硅谷已写好了全球AI竞赛的剧本:闭源模式,限制技术扩散;算力堆砌,抬 高追赶壁垒;垄断优势,获得高昂商业利润。 "DeepSeek、通义千问等一批大模型加速发展",写入了2025年的杭州市政府工作报告中。以低成本打破赛道壁垒、震动全球同业的DeepSeek开源大模型背 后,是创新活力的迸发。杭州是如何发展开源大模型的,"开源大模型之城"为什么是杭州? 随着DeepSeek以开源模式引发行业变革,开源迅速成为大模型主流开发模式。 4月2日,全球最大AI开源社区HuggingFace发布最新榜单,排在前三的开源大模型分别来自阿里通义千问、DeepSeek和群核科技,领先于英伟达、谷歌等 公司。 榜单发布后,杭州再次引起业界瞩目。因为杭州包揽了前三,成为全球少有的、同时拥有3个世界顶级开源模型的城市,因此被誉为"开源大模型之城"。 开源大模型对AI普及应用、构建AI产业生态至关重要。目前,北京等地都在积极打造"全球开源之都",而杭州走在了前列。 杭州"开源大模型之城"是如何炼成的? 01 深厚土壤 然而,DeepSeek反其道而行之,凭借开源和低成本 ...
早报|特朗普称哈佛大学国际生比例最高15%;泡泡玛特回应Labubu品控问题;苹果计划全面重命名操作系统;荣耀回应机器人业务
虎嗅APP· 2025-05-28 23:55
Group 1: Education and International Relations - The U.S. government is imposing restrictions on Harvard University regarding international students, suggesting a cap of 15% on foreign students, which currently stands at approximately 31% [2] - The U.S. government has also announced the cancellation of federal funding for Harvard and has suspended new student visa interviews [2] Group 2: Financial Services and Investment - Chinese Vice Premier He Lifeng met with Morgan Stanley's co-president, expressing a commitment to high-level openness and inviting more U.S. financial institutions to deepen cooperation in China's capital market [3] - The Chinese Foreign Ministry emphasized that the essence of Sino-U.S. economic relations is mutual benefit, highlighting the significant bilateral demand reflected in increased orders from U.S. buyers [4] Group 3: Consumer Goods and Quality Control - Pop Mart's Labubu plush toys have gained popularity, but there are reports of quality control issues, including defects like misalignment and paint loss, leading to customer dissatisfaction [6] - Pop Mart's customer service stated that all products undergo quality checks before shipment, but minor imperfections may occur during production [6] Group 4: Technology and Innovation - Didi Enterprise Edition has become the first travel service provider for 3M in China, offering innovative services that have led to a 39% year-on-year increase in ride orders from foreign clients [8] - DeepSeek has released an open-source version of its R1 model, which reportedly performs comparably to OpenAI's latest models [9] Group 5: Pharmaceuticals and Healthcare - Fosun Pharma has signed exclusive commercialization agreements for several biopharmaceutical products with Nine Sources Gene, covering regions including the Middle East and parts of Southeast Asia [17] - The National Healthcare Security Administration is conducting checks on retail pharmacies to address potential issues of pharmacists' credentials being misused [16]
78%主创跳槽,Llama 14名作者只剩3人,Meta最强开源模型团队大溃散引争议
3 6 Ke· 2025-05-27 12:19
AI 人才争夺战愈演愈烈,就算是顶级大厂,如果没有"护城河",也留不住人。 据外媒 Business Insider 最新消息,曾在开源大模型圈子里一度领跑的 Meta,如今正面临严重的人才流失。在 Llama 模型最初的 14 位核心作者 中,已有 11 位离职。有的自立门户,有的跳槽去了竞争对手。 这波"出走潮"也让外界再次把目光投向 Meta。毕竟他们曾豪赌元宇宙,四年"烧掉"450 亿美元,却被直指至今几乎未见显著成效。现在 AI 项目 也出问题了,不少人开始质疑:Meta 还行不行?为什么留不住顶尖 AI 人才?它的创新能力,还能支撑它在这场 AI 竞赛中跑多远? Llama 论文的 14 位作者,已有 11 人离开 Meta 回头看 2023 年那篇引发轰动的 Llama 论文,共署名 14 位研究者。短短两年,Meta 只留下了其中三位:研究科学家 Hugo Touvron、研究工程师 Xavier Martinet 和项目负责人 Faisal Azhar。 论文地址:https://arxiv.org/pdf/2302.13971 其他 11 人,大多已经离开,分散到了全球多家科技公司,有的还 ...
夸克升级“深度搜索”功能,AI应用方向催化丰富,关注影视、游戏景气度回暖
Huachuang Securities· 2025-05-12 00:15
Investment Rating - The report maintains a "Recommendation" rating for the media industry, expecting the industry index to rise more than 5% over the next 3-6 months compared to the benchmark index [49]. Core Insights - The media sector is experiencing a resurgence in the film and gaming markets, driven by advancements in AI applications and cultural confidence stemming from popular IPs like "Nezha" [3][6]. - The report highlights the importance of AI applications in reshaping the industry landscape, with a focus on public cloud value reconstruction and the return to growth for related companies [6][9]. - The gaming market shows strong performance, with Tencent and NetEase leading in iOS sales rankings, indicating a healthy competitive environment [16][19]. - The film market is recovering, with total box office revenue reaching 239 billion yuan as of May 9, 2025, which is approximately 102% of the 2019 level [20][23]. Market Performance Review - The media (Shenwan) index rose by 1.40% last week, underperforming the CSI 300 index, which increased by 2.00%, resulting in a relative underperformance of 0.61% [9][10]. - The report notes that the media sector ranks 24th among all sectors in terms of performance [9]. Industry Highlights - The report emphasizes the potential of AI applications in various sectors, including gaming and education, suggesting a focus on companies like Huya, Giant Network, and Perfect World for gaming, and New Oriental and TAL Education for education [6][30]. - The film industry is expected to benefit from a strong pipeline of upcoming releases, with several key films set to debut in mid-May [29][30]. Company Announcements - Perfect World announced a 2025 employee stock ownership plan, indicating a commitment to employee engagement and retention [36]. - Wanda Film disclosed plans for a share reduction by a major shareholder, which may impact market sentiment [38].
贸易战下的产业韧性(二):AI大模型的商业“回旋镖”,重新落到了云计算
3 6 Ke· 2025-05-11 23:28
Core Viewpoint - The domestic large model industry is attempting to break through its current challenges and reconstruct a new order, but the unstable market environment poses significant risks [1] Group 1: Open Source Trends - DeepSeek has disrupted the industry's perception of open-source models, prompting OpenAI's CEO to reconsider the validity of open-source strategies [1] - Domestic large model companies like Alibaba, Baidu, and SenseTime are accelerating their open-source initiatives [1] - Open-source is seen as a key strategy to reduce dependency on foreign software and hardware, but the commercial viability of open-source projects remains complex [2][5] Group 2: Challenges in Implementation - Developers face significant technical adaptation and maintenance costs, despite open-source models lowering the technical barrier [4] - The integration of large models into existing systems requires extensive customization, which can be resource-intensive for companies [4] - The complexity of data acquisition, cleaning, and labeling poses additional challenges for businesses, particularly small and medium-sized enterprises [4] Group 3: Investor Sentiment - Investors are cautious about the open-source model due to the unclear profitability and traditional software sales evaluation methods not being applicable [5] - The potential for significant financial loss if investments in proprietary models are undermined by open-source alternatives is a concern for investors [4][5] Group 4: Business Models - Chinese large model companies are adopting a "free-to-use plus value-added services" model to build a commercial framework around open-source models [6][8] - Companies like Baidu are leveraging their cloud services to monetize the usage of their open-source models, creating a win-win situation for developers and the company [8] - The success of open-source models may depend more on the quality of cloud services than on the models themselves, as seen in the strategies of Meta and Hugging Face [9][10] Group 5: Future Outlook - Open-source is viewed as a pathway for the Chinese large model industry to overcome technological barriers, but commercial sustainability is equally important [10] - The increasing tariff barriers from the U.S. add pressure to the large model industry, making the choice of cloud platforms more critical than the open-source models themselves [10]