作为"AI六小虎"之一的MiniMax（稀宇科技），近期刚刚结束了"技术发布周"。 自6月17日以来，MiniMax发布了一系列核心技术与产品更新，涵盖基座模型、视频生成模型、多模态智能体等多个方向，系统展现其"模型--多模态--应 用"一体化技术路线。 就在MiniMax"疯狂上新"的同时，市场上传出了其拟赴港IPO的消息，这也让MiniMax成了智谱AI之后第二家"AI六小虎"中传出冲刺上市的公司；不过， 目前公司对此没有相关回应。 图源：微博截图 即使身处时下最火的赛道、顶着明星履历，MiniMax或者其他"AI六小虎"的企业也都面临着同样的困境——产品同质化、商业化不成熟，以及盈利困境。 01.冲刺IPO的"六小虎" "AI 六小虎"（谱、月之暗面、百川智能、MiniMax、阶跃星辰、零一万物）是过去两年国内大模型时代的一个标志，指的是当年最早完成 10 亿 + 美元融 资，且均拥有自研千亿参数级大模型，在国际基准测试中与 GPT-4、Llama 等对标的大模型创业公司。 为了避免依赖单一产品的风险，2024年9月，MiniMax发布了视频生成模型"海螺视频I2V-01"，并上线了海螺AI。 AIGC ...

文 | 新立场Pro 25年刚过半，"AI六小虎"的称呼仿佛成为了一个过气梗，DeepSeek横空出世，改写了AI创业的造神势头——原本也许可以在新一年高歌猛进的六大玩 家，上半年有人削减投流，有人裁撤业务，有人重心转向，有人更加急切地寻求上市。 2月消息称月之暗面Kimi暂停烧钱投流；3月消息称百川智能裁撤主要负责金融行业To B业务的B端组；4月智谱向中国证监会北京证监局提交上市辅导备 案；5月零一万物转向彻底拥抱To B；6月初消息称阶跃星辰减少C端应用投入并将重心转向Agent；就在前段时间，消息称MiniMax计划最早在今年港股上 市，并且已为此次IPO聘请了财务顾问。 DeepSeek鲶鱼效应加速下，六小虎当然知道市场留给自己的耐心更加不多了，寻求上市则是赚取更多市场耐心的方式之一。 不止闫俊杰，根据公开资料，MiniMax另一位创始人周彧聪也来自商汤。所以起初，在市场对AI六小虎的模式和演变路径都还不清楚的时候，MiniMax被 视为"商汤系"、"商汤2.0"，这当然不意味着公司业务层面在效仿商汤的模式，仅仅指向管理层的人员履历。 以近年对外发言较多的闫俊杰的基本履历来看，其此前其在商汤科技负责 ...

Core Viewpoint - The article discusses the transformative impact of large models in the tech industry, emphasizing that innovation is the key survival strategy for companies in this space, especially in light of the rapid evolution and competition among startups and tech giants [2][3][14]. Group 1: Industry Trends - The large model industry is experiencing a significant shift towards innovation, with traditional internet business models becoming obsolete [3][4]. - The recent "Aha Moment" in the industry, exemplified by viral videos of animals performing complex actions, highlights the advancements in video AI technology and its potential [7][8]. - The MiniMax Week event serves as a critical point for examining how startups can thrive amidst competition from larger firms [4][6]. Group 2: Technological Innovations - MiniMax's Hailuo 02 model has seen a threefold increase in parameters compared to its predecessor, achieving native 1080P resolution and generating 10 seconds of high-definition content [9][10]. - The model's innovative NCR architecture allows for efficient resource allocation, significantly reducing memory read/write by over 70% and improving training and inference efficiency by 2.5 times [12][23]. - MiniMax's M1 model, with 456 billion parameters, supports the longest context length in the industry, enhancing its performance in complex tasks [16][18]. Group 3: Competitive Landscape - Despite the initial dominance of tech giants in the large model space, startups like MiniMax have captured significant market share and achieved top rankings in performance benchmarks [15][16]. - The article notes that the rapid evolution of large models requires companies to continuously innovate to maintain a competitive edge, as capital alone is insufficient for success [14][15]. - MiniMax's innovative approaches, such as the use of mixed attention mechanisms and the CISPO training method, have allowed it to outperform competitors while reducing costs [20][21][23]. Group 4: Agent Applications - The emergence of agent applications, such as MiniMax Agent, represents a new frontier in AI, enabling more complex task execution and planning capabilities [30][32]. - MiniMax Agent has been integrated into daily operations, demonstrating its effectiveness in various tasks, including programming and content creation [31][32]. - The synergy between large model innovations and agent applications is expected to drive further growth and development in the AI ecosystem [32][34].

Core Viewpoint - MiniMax has launched a new AI model, M1, claiming its performance can compete with top models from OpenAI, Anthropic, and Google DeepMind, while its training and operational costs are significantly lower [1][3]. Group 1: Model Performance and Cost - MiniMax's M1 model reportedly matches the intelligence and creativity of leading models but was trained at a cost of only $534,700, which is nearly 200 times lower than the estimated training cost of ChatGPT-4o, potentially exceeding $100 million [3]. - The introduction of M1 could disrupt the market demand for OpenAI's products, as OpenAI has already begun to reduce prices for its models to maintain market share [4]. Group 2: Market Impact and Industry Reactions - If M1's performance is validated, it may affect the profitability of cloud service providers like Amazon AWS, Microsoft Azure, and Google Cloud, as companies may not need to invest heavily in computational resources [5]. - The announcement of M1 has not yet caused significant market fluctuations, unlike the previous launch of DeepSeek's R1 model, which led to a 17% drop in Nvidia's stock price [5]. Group 3: Model Features and Accessibility - M1 features a context window of 1 million tokens, allowing it to process more data than some leading models, such as OpenAI's o3 and Anthropic's Claude Opus 4, which have context windows of approximately 200,000 tokens [7][8]. - Users can access M1 for free through its API, and developers can download the entire model to run on their own resources, which may enhance its adoption [7]. Group 4: Company Background and Support - MiniMax is backed by major Chinese tech companies like Tencent and Alibaba, although details about its employee size and CEO are limited [6].

华夏时报（www.chinatimes.net.cn）记者 石飞月 北京报道 作该公司对标Manus的产品；6月20日，Hailuo Video Agent亮相，这是一款视频创作Agent。 这其中会不会产生下一款爆品，现在还未可知，但支柱产品单一对企业来说本身就有很大的市场风险，再加上 Talkie去年还经历了一次下架风波，因此MiniMax扩大产品线的布局是非常必要的。 根据Sensor Tower去年7月的数据，当时Talkie全球月活数据已达到1100万，超过一半用户来自美国，其他市场包括 菲律宾、英国以及加拿大等。然而在去年12月，据媒体报道，Talkie经历了一次被苹果商店下架的风波。对于下架 的原因《华夏时报》记者采访了 MiniMax方面，对方未就此做出回复。 去年12月，OpenAI完成了连续12个工作日的12场直播，本周，"AI六小虎"之一的MiniMax（稀宇科技）效仿 OpenAI，上演了一次新品"闪电战"。截至6月21日，该公司本周已推出开源大规模混合架构推理模型MiniMax- M1、新视频生成模型 Hailuo 02、MiniMax Agent和一款视频创作Agent，预计下周一 ...

MiniMax上演逆袭。 作者 I 刘博 报道 I 投资界PEdaily 全球AI圈又沸腾起来了。 6月17号，Mi n iMa x "发布周"拉开序幕，又是开源又是发布新模型， 一系列动作令业内 惊呼。 新模型Mi niMa x-M1，是"世界上第一个开源的大规模混合架构的推理模型"，被 视为"性价比新王"。值得一提的是，这只是一道"开胃菜"。 Mi n iMa x，这个名字创投圈并不陌生。身后掌舵者是一位8 9年的博士——闫俊杰，他早 年任职于AI四小龙之一的商汤科技，后来在2 022年投身到AGI创业之中。早在Ch a tGPT 爆红前，Mi n iMa x就已经成为国产大模型的代表性企业。 1 9 8 9年，闫俊杰出生在河南的一座小县城。他从中国科学院自动化研究所博士毕业后， 继续在清华大学计算机系从事博士后研究。他的学术成就令人印象深刻——在深度学习和 计算机视觉领域，发表过顶级会议和期刊论文1 0 0余篇, Go ogl e Sc h o l a r引用超过10 0 0 0 次。 博士期间，闫俊杰加入商汤科技，从实习生一路做起，曾任商汤集团副总裁、研究院副院 长 和 智 慧 城 市 事 业 群 C ...

Core Insights - The competition among the "Six Little Tigers" (MiniMax, Zhipu, Moonlight, Baichuan Intelligence, Zero One Everything, and Jiyue Star) is intensifying as they strive to prove their capabilities against DeepSeek, particularly in the development of reasoning models [1][3] - MiniMax has launched several new products, including the M1 reasoning model and the MiniMax Agent, as part of its strategy to remain competitive and relevant in the market [3][4] - The IPO ambitions of the "Six Little Tigers" are facing challenges due to revenue requirements and market conditions, with only Zhipu currently meeting the necessary financial criteria [9][11] Group 1: Product Development and Competition - Moonlight and Zhipu have released reasoning models that compete with DeepSeek's R1, with Moonlight's Kimi-Dev-72B model outperforming R1 in AI programming tests despite having significantly fewer parameters [1][3] - MiniMax's M1 model supports 1 million context inputs, which is eight times that of R1, marking a significant technological advancement [3] - MiniMax's recent product launches include the M1 model, video generation model Hailuo 02, and the MiniMax Agent, indicating a strategic shift towards diversifying its product offerings [4][5] Group 2: Market Position and IPO Aspirations - MiniMax's revenue has historically relied on its flagship product, Talkie, which has faced challenges, including a temporary removal from app stores [4][12] - The company is expanding its revenue streams by introducing new products like Hailuo AI and MiniMax Agent, targeting higher-paying overseas markets [12] - The IPO landscape for the "Six Little Tigers" is complicated, with only Zhipu having submitted its listing application, while MiniMax is still preparing its IPO materials amid challenging market conditions [9][10][13]

自6月17日起，MiniMax将连续发布5天新模型。 本文为IPO早知道原创 作者｜ Stone Jin 微信公众号｜ipozaozhidao 值得注意的是，自 6月17日起， MiniMax 将连续发布 5天新模型。 其中， 6月17日发布的 MiniMax-M1 是 世界上第一个开源的大规模混合架构的推理模型 —— M1 在面向生产力的复杂场景中能力是开源模型中的最好一档，超过国内的闭源模型，接近海外的最领先 模型，同时又有业内最高的性价比。 与 DeepSeek-R1相比，M1模型在数学和编码等任务中效率更高 ；与 GPT-4 相比， M1推理模 型 能力与之匹配 ， 而 成本仅为 GPT-4的0.5% 。 6月18日， MiniMax 又发布了 新视频生成模型 Hailuo 02 ，同样 打破全球视频模型效果成本纪录 。 本文由公众号IPO早知道（ID：ipozaozhidao）原创撰写，如需转载请联系C叔↓↓↓ 据 IPO早知道消息，国内"大模型六小龙之一"的 MiniMax （稀宇科技）最快今年赴港上市。 一操办过多起香港大型科技 IPO的投行人士向「IPO早知道」透露， MiniMax 的确就赴 ...

Core Viewpoint - Lexibook reported strong annual financial results for the fiscal year ending March 31, 2025, with significant growth in revenue, operating income, and net income, driven by successful product launches and digital sales, but faces uncertainties for the upcoming fiscal year due to potential increases in U.S. customs duties and other external factors [3][20][21]. Financial Performance - Net sales reached €74.0 million, a 26.3% increase from €58.6 million in the previous year [2][3]. - Gross margin improved to €44.6 million, up 27.0% from €35.1 million, with a gross margin rate of 60.2% compared to 59.9% [2][8]. - Operating income rose to €8.1 million, a 36.1% increase from €6.0 million, while EBITDA increased by 30.3% to €9.7 million [2][10][13]. - Net income for the year was €7.4 million, up 47.7% from €5.0 million [2][16]. Strategic Developments - The Group experienced six consecutive years of growth, attributed to successful strategic choices, high-potential new products, and strong digital sales [5][6]. - The "Stitch effect" contributed €17 million in revenue, highlighting the impact of licensed products on overall sales [5][6]. - A significant digital advertising campaign generated over one billion impressions, enhancing brand awareness [11]. Market Dynamics - France accounted for 34% of total revenue, while international sales outside Europe increased from 10% to 13% [6]. - The Group's digital sales are growing rapidly, supported by a large-scale global marketing campaign [7]. Financial Position - The net cash position improved to €7.1 million from €2.0 million, primarily due to cash generation and repayment of non-current liabilities [18]. - Equity reached a historic high of €24.4 million, reflecting the Group's strengthened financial position [3][4]. Outlook and Challenges - The outlook for the 2025-26 fiscal year is uncertain due to potential increases in U.S. customs duties, which could significantly impact growth and margins [20][21]. - The Group anticipates a strong start to the fiscal year, driven by the release of the Stitch movie, but acknowledges risks related to trade relations and competition [19][21].

闫德利 腾讯研究院资深专家 一、人工智能是国家战略 我国高度重视人工智能的创新和发展。早在2014年，习近平总书记在中国科学院第十七次院士大会、中 国工程院第十二次院士大会上就对人工智能进行了强调。2017年，人工智能首次写入《政府工作报 告》；同年国务院印发《新一代人工智能发展规划》，提出目标"到2030年人工智能理论、技术与应用 总体达到世界领先水平"。次年，首届世界人工智能大会在上海召开。此后，人工智能在众多重要会议 上不断得以强调。尤其是今年，中共中央政治局就人工智能进行了集体学习，习近平总书记调研了上 海"模速空间"大模型创新生态社区，《政府工作报告》大篇幅描述人工智能，河南、福建、河北、辽 宁、山东、北京、江西、天津、新疆、陕西等地党委理论学习中心组纷纷就人工智能组织学习会。我国 掀起了 人工智能研发和应用的热潮 。如下表所示。 | 会议 | 日期 | 主题/相关表述 | | --- | --- | --- | | 政治局集体 | 2018.10 | 人工智能发展现状和趋势 | | 学习 | 2025.4 | 加强人工智能发展和监管 | | 政治局会议 | 2023.4 | 要重视通用人工智能发展 ...