[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"$fjP_UI_8sAA6MHpTGHJRKca4-ExejXiEC5H_mpZFHHpk":3},{"code":4,"msg":5,"data":6},200,"操作成功",{"id":7,"title":8,"content":9,"digest":10,"source":10,"coverPath":11,"thumbsCoverPath":12,"isTop":13,"isShow":14,"baseClick":13,"clickCount":15,"createTime":16,"typeId":17,"isNewest":18,"newsInfoTypeRespVo":19,"voiceUrl":22,"voiceSize":23,"taskId":24,"releaseTime":25,"titleEn":26,"contentEn":27,"voiceUrlEn":28,"taskIdEn":29,"voiceSizeEn":30},1356,"672个AI“选手”集结！中国生成式AI备案登记图鉴，教育AI竟成“卷王”","\u003Cp>\u003Cstrong class=\"ql-lineHeight-1-75\" style=\"font-size: 18px; color: rgb(255, 153, 0);\">得益于AI技术的成熟和开源大模型的发展，中国生成式AI应用及服务正迎来爆发式增长。\u003C\u002Fstrong>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">根据国家网信办公示的生成式AI服务备案信息：\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">截至2025年6月，共439款生成式人工智能服务在国家网信办完成备案；\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">通过API接口或其他方式直接调用已备案模型能力的AIGC应用或产品完成登记的有233个（大部分产品表现形态为Agent）。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">简单来说，如果是自研或者有显著微调、要做数据训练的大模型，需要备案。如果是基于已备案的大模型、无微调也无数据训练而开发的AIGC应用产品，无需备案，只需登记。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">IT桔子对这些备案和登记的中国生成式AI模型及其应用的数据做进一步地处理，包括详细记录类型（通用或垂直）、应用场景、备案\u002F登记单位的行业及背景、融资情况等关键维度。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">本文将基于上述备案\u002F登记信息进行详细解读。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cstrong class=\"ql-lineHeight-1-75\" style=\"font-size: 18px; color: rgb(255, 153, 0);\">一、增长态势：AIGC大模型2024年迎来爆发拐点\u003C\u002Fstrong>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cimg alt=\"undefined\" src=\"https:\u002F\u002Fimage.51xinwei.com\u002F2025\u002F09\u002Fc515349d1f1844a3a9f70604e290842b\u002F744.jpg\" width=\"undefined\" height=\"undefined\" style=\"display: block; margin: auto;\">\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">2023年8月国家网信办发布《生成式人工智能服务管理暂行办法》，当年即有61项生成式大模型完成备案，标志着生成式AI在中国的起步阶段。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">这一时期以企业自研的通用模型为主，如百度“文心一言”和阿里“通义千问”，得益于ChatGPT全球热潮的刺激，中国企业快速响应，推出基础大模型。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">2024年，国内生成式AI大模型备案数量爆发式增长至241项，同比增长约3倍。这一年备案数量激增，反映了政策窗口期以及基础大模型技术成熟的双重因素推动。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">同时，基于通用模型能力的而产生的大量AIGC 产品\u002F模型蓬勃兴起，相应的服务快速推出，登记数量显著增加。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">2025年上半年，生成式AI大模型备案已达137项，增速虽放缓但仍强劲。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cstrong class=\"ql-lineHeight-1-75\" style=\"font-size: 18px; color: rgb(255, 153, 0);\">二、技术路线演进：从通用智能到场景智能\u003C\u002Fstrong>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">中国AIGC大模型生态正趋向多元化，并走出了从通用——垂直——场景的演进路线。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cimg alt=\"undefined\" src=\"https:\u002F\u002Fimage.51xinwei.com\u002F2025\u002F09\u002F554a4de9c3db457f8ef107a65bffaf98\u002F746.jpg\" width=\"undefined\" height=\"undefined\" style=\"display: block; margin: auto;\">\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">据IT桔子统计，在2024年之前，通用大模型备案登记数量较多，典型代表如百度文心一言、智谱ChatGLM等，强调多模态能力和广泛适用性，适合跨行业应用。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">然而，基础大模型在解决特定行业的痛点时仍存在不足，需要结合行业知识、经验、数据优化，近两年专注特定领域的垂直大模型数量也在快速增加，并在2025年上半年超越了通用大模型，成为备案的主流。垂直大模型包括教育领域的“EduChatT”，医疗领域的西湖大模型，财务领域的金蝶苍穹大模型，法律领域的中国-东盟法律大模型等。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">与此同时，2024下半年基于大模型而构建的具体细分场景应用产品（比如客服）实现了快速增长，从13个登记量增长到89个，而且2025年上半年已登记的Agent产品有127个，超越去年全年的数据。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">这些数据反映了快速增长的市场需求，体现了AI从“广度”向“深度”转型的趋势，表明市场正在从\"技术炫技\"转向\"场景深耕\"。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cstrong class=\"ql-lineHeight-1-75\" style=\"font-size: 18px; color: rgb(255, 153, 0);\">三、场景应用图谱：教育医疗双擎驱动\u003C\u002Fstrong>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">据IT桔子统计，国内AIGC大模型的主要应用场景分布包括教育、医疗、金融、客服、工业等各个领域，涉及办公、购物、娱乐、学习等方方面面，显示出AI向实体经济的渗透。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cimg alt=\"undefined\" src=\"https:\u002F\u002Fimage.51xinwei.com\u002F2025\u002F09\u002F5809bf8eca324c9fb2da1602ddafea80\u002F749.jpg\" width=\"undefined\" height=\"undefined\" style=\"display: block; margin: auto;\">\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">其中，教育AIGC大模型及其相关应用共有39项，总数位列第一，遥遥领先于其他场景，代表有网易有道的“子曰”大模型、好未来MathGPT大模型和猿力科技的“看云大模型”。AI在个性化教学、口语练习等具体的教育场景中大放异彩，响应了“双减”政策和在线教育需求。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">医疗大模型及相关AIGC应用产品共有21项，总数第二，包括小荷健康发布的小荷AI医生、好心情健康产业发布的“好心情AI陪伴”等，聚焦诊断辅助、患者咨询和心理疗愈。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">像工商银行等各大企业还通过大模型能力研发专注于客服场景的AIGC产品，共有16项，用于提升智能客服咨询效率，常见于电商和金融领域。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">金融大模型代表有蚂蚁金服发布的AntGLM大模型和度小满轩辕大模型、同花顺的问财HithinkGPT大模型等，应用于风控、理财咨询。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">基于大模型能力，面向文旅行业的AIGC应用服务也有不少，如马蜂窝的“AI行程助手”和“滴滴差旅助手”，优化旅游规划和出行服务。此外，携程还发布了“问道”垂直文旅大模型，同程旅行发布了“程心大模型”。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">媒体行业也有Agent智能产品，比如财联社小财神，用于AI内容生成、信息解读。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">这些场景侧面反映了中国AI的“实用主义”导向。未来，随着多模态AI（如文生图、语音交互、视频生成）成熟，跨场景融合将成为趋势，推动AI从单一应用向生态系统演进。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cimg alt=\"undefined\" src=\"https:\u002F\u002Fimage.51xinwei.com\u002F2025\u002F09\u002F8e290550b0b84e269ec5f60abd66eafa\u002F751.jpg\" width=\"undefined\" height=\"undefined\" style=\"display: block; margin: auto;\">\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cstrong style=\"font-size: 18px; color: rgb(255, 153, 0);\">四、地域分布：北京全面领跑，上海、广东紧随其后\u003C\u002Fstrong>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">中国AI备案的加速与监管政策、各地的创业创新环境密切相关。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">2023年作为生成式AI大模型的“元年”，备案主要集中在北京、上海等一线城市；\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">2024-2025年则向全国扩散，广东、江苏、浙江、四川等省份一些基于大模型衍生的AIGC的产品增多。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cimg alt=\"undefined\" src=\"https:\u002F\u002Fimage.51xinwei.com\u002F2025\u002F09\u002F56e2ad409dfa48b68c47acd9605435c5\u002F755.jpg\" width=\"undefined\" height=\"undefined\" style=\"display: block; margin: auto;\">\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">具体来看，目前北京市的大模型及AIGC产品最多，约占总量的31%，其中，已登记的Agent产品有73个，而已备案的大模型数量就达到了132个，以通用大模型为主，如百度“文心一言”、出门问问的“序列猴子”大模型、美团大模型“通慧”等。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">这充分体现了北京企业具有领先的AI大模型基础研发能力，技术前瞻性走在了前列，主要得益于首都的科研资源（如中科院）和互联网巨头聚集。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">上海市以26%的占比紧随其后，其中，已登记的Agent产品有95个，已备案的大模型数量有82个，如商汤的“日日新”大模型、上海人工智能实验室发布的书生·浦语大模型、阶跃星辰发布的阶跃等。说明上海企业的大模型研发实力不容小觑，但AIGC产品的落地应用响应速度更强，更具商业化思维。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">广东约占11%，其中已登记的Agent产品有8个，已备案的大模型数量有66个，除了腾讯的混元助手大模型、元象XVERSE大模型外，其余主要是硬件厂商、电子公司发布的，如vivo的“蓝心大模型”、OPPO的AndesGPT-LVM、荣耀的魔法大模型、中兴通讯的星云大模型、TCL大模型助手等。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">其他省份如江苏、浙江、贵州等各地也有登记的AIGC产品，体现区域均衡发展。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cstrong class=\"ql-lineHeight-1-75\" style=\"font-size: 18px; color: rgb(255, 153, 0);\">五、研发主体生态：多层次创新梯队\u003C\u002Fstrong>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">IT桔子对大模型的研发主体进行分类统计，结果如下：\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cimg alt=\"undefined\" src=\"https:\u002F\u002Fimage.51xinwei.com\u002F2025\u002F09\u002Fe2c59d2f45a54f7999f3b3643c81bd96\u002F760.jpg\" width=\"undefined\" height=\"undefined\" style=\"display: block; margin: auto;\">\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">有近一百家上市公司是大模型的参与主体，如百度文心一言、阿里通义、腾讯混元大模型，以及科大讯飞、昆仑万维等。它们凭借资金和技术优势，主导通用大模型的开发。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">有47项AIGC大模型是由国内独角兽企业独立研发，如智谱AI发布的多款大模型——智谱清言、清影AI、智谱端侧，抖音发布的云雀大模型、百川智能的百应大模型、MiniMax的abab模型，月之暗面发布的Moonshot大模型等，多为还未上市的高估值企业。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">除独角兽外的大众创业公司也参与了大模型或Agent产品研发，体现创新活力。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">IT桔子注意到，虽然民营科技公司是AI大模型的主要参与者，但包括央企国企（中国移动九天大模型、中石油昆仑大模型）、高校及科研机构（中科院自动化所的紫东太初大模型）和医院等其他组织主体亦未缺席这场赛事。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">据统计，目前参与的高校已有：哈尔滨工业大学人工智能研究院发布的天工开悟大模型，同济大学发布的CivilGPT，以及复旦大学附属妇产科医院发布的Agent智能体产品“小红AI患者助理”，上海交通大学医学院附属瑞金医院发布的Agent产品“瑞金医院医学智能问答”，南京鼓楼医院的Agent产品“骨灵”等。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">这一结构凸显中国AIGC的“多层次”生态：上市公司提供稳定性，创业公司注入活力，科研机构推动基础创新。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">综上所述，中国生成式AI正进入“黄金时代”：备案登记总量从2023年的61项飙升至2025上半年的累计672项，增长超10倍，体现了国家战略（如“新一代人工智能发展规划”）的成效。垂直模型的崛起标志AI从“通用工具”向“行业赋能”转型，尤其在教育、医疗等应用领域AIGC产品发展迅速，体现了AI的实用价值导向。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cspan style=\"color: rgb(187, 187, 187);\">【新闻来源】 IT桔子 \u003C\u002Fspan>\u003Cspan style=\"color: rgb(187, 187, 187); background-color: rgb(56, 56, 56);\">作者| 吴梅梅  \u003C\u002Fspan>\u003Ca href=\"https:\u002F\u002Fmp.weixin.qq.com\u002Fs?__biz=MjM5MjQ2NzA2Mg==&mid=2649624127&idx=1&sn=179661e187cf3f2ab004fdb1da39e9be&poc_token=HHbaw2ijGb0jU2lDAxKWi8UP_kaN_URBWsKQbYcT\" rel=\"noopener noreferrer\" target=\"_blank\" style=\"color: rgb(187, 187, 187);\">http:\u002F\u002Fi9n.cn\u002FcvDgK\u003C\u002Fa>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cspan style=\"color: rgb(187, 187, 187);\">（本网转发此文章，旨在为读者提供更多的信息资讯，所涉内容不构成投资、消费建议。文章事实如有疑问，请与有关方核实，文章观点非本网观点，仅供读者参考。）\u003C\u002Fspan>\u003C\u002Fp>","","https:\u002F\u002Fimage.51xinwei.com\u002F2025\u002F09\u002F161a4d40302a467ba5a48ca70bec8525\u002FAI领域.jpg","https:\u002F\u002Fimage.51xinwei.com\u002F2025\u002F09\u002Fthumbs\u002F161a4d40302a467ba5a48ca70bec8525\u002FAI领域.jpg",0,1,42,"2025-09-12 16:43",2,false,{"id":17,"name":20,"enName":21},"芯位视野","Xinwei Vision","https:\u002F\u002Fxinwei-dev-test.oss-cn-shenzhen.aliyuncs.com\u002Fintelligent\u002Faudio%3Aad775d34-e081-47e1-a02d-8a909b91a080%3A0.wav?Expires=1757676252&OSSAccessKeyId=LTAI5tNvY2RkKjZw4LLWsrPK&Signature=TTWVQ2iT1%2FCgIw2Wrzcv07OqrnM%3D",16545440,"ad775d34-e081-47e1-a02d-8a909b91a080","2025-09-12 16:29","672 AI \"competitors\" have gathered! China's Generative AI Filing Guide, Education AI becomes the \"king of competition\"","\u003Cp>\u003Cstrong class=\"ql-lineHeight-1-75\" style=\"font-size: 18px; color: rgb(255, 153, 0);\">Thanks to the maturity of AI technology and the development of open-source large models, China's generative AI applications and services are experiencing explosive growth.\u003C\u002Fstrong>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">According to the publicly disclosed generative AI service filing information from the Cyberspace Administration of China:\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">As of June 2025, a total of 439 generative artificial intelligence services have been filed with the Cyberspace Administration of China;\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">There are 233 AIGC applications or products that have completed registration by directly calling the capabilities of the registered models through API interfaces or other methods (most of the products are in the form of Agent).\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">In simple terms, if it is self-developed or has significant fine-tuning and requires data training, it needs to be filed. If it is a AIGC application product developed based on an already filed large model without any fine-tuning or data training, it does not need to be filed, only registered.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">IT桔子 further processes the data of these Chinese generative AI models and their applications that have been filed and registered, including detailed records of types (general or vertical), application scenarios, industry and background of the filing\u002Fregistration units, financing situations, and other key dimensions.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">This article will provide a detailed interpretation based on the above filing\u002Fregistration information.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cstrong class=\"ql-lineHeight-1-75\" style=\"font-size: 18px; color: rgb(255, 153, 0);\">I. Growth Trend: The AIGC Large Model Achieved a Breakthrough Point in 2024\u003C\u002Fstrong>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cimg alt=\"undefined\" src=\"https:\u002F\u002Fimage.51xinwei.com\u002F2025\u002F09\u002Fc515349d1f1844a3a9f70604e290842b\u002F744.jpg\" width=\"undefined\" height=\"undefined\" style=\"display: block; margin: auto;\">\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">In August 2023, the Cyberspace Administration of China issued the \"Interim Measures for the Management of Generative AI Services\", and in the same year, 61 generative large models were filed, marking the initial stage of generative AI in China.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">During this period, the main focus was on general models developed by enterprises, such as Baidu's \"Wenxin Yinyi\" and Alibaba's \"Tongyi Qianwen\", which benefited from the global popularity of ChatGPT, prompting Chinese companies to quickly respond and launch basic large models.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">In 2024, the number of large model filings for generative AI in China surged to 241, a threefold increase compared to the previous year. This sharp increase in filings reflected the dual factors of the policy window period and the maturity of basic large model technology.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">At the same time, a large number of AIGC products\u002Fmodels emerged based on the capabilities of general models, and corresponding services were rapidly launched, resulting in a significant increase in registration numbers.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">In the first half of 2025, the number of large model filings for generative AI had reached 137, with a slower growth rate but still strong.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cstrong class=\"ql-lineHeight-1-75\" style=\"font-size: 18px; color: rgb(255, 153, 0);\">II. Technological Evolution: From General Intelligence to Scenario Intelligence\u003C\u002Fstrong>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">The Chinese AIGC large model ecosystem is becoming more diverse, and has taken a path from general - vertical - scenario.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cimg alt=\"undefined\" src=\"https:\u002F\u002Fimage.51xinwei.com\u002F2025\u002F09\u002F554a4de9c3db457f8ef107a65bffaf98\u002F746.jpg\" width=\"undefined\" height=\"undefined\" style=\"display: block; margin: auto;\">\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">According to IT Juzi statistics, before 2024, the number of general large model filings and registrations was higher, with typical representatives such as Baidu's Wenxin Yinyi and Zhipu ChatGLM, emphasizing multimodal capabilities and broad applicability, suitable for cross-industry applications.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">However, basic large models still have shortcomings in solving specific industry pain points, requiring integration of industry knowledge, experience, and data optimization. In recent years, the number of vertical large models focusing on specific fields has also increased rapidly, and in the first half of 2025, they surpassed general large models and became the mainstream of filings. Vertical large models include \"EduChatT\" in the education field, the West Lake Large Model in the medical field, the Kingdee Cangqiong Large Model in the financial field, and the China-ASEAN Legal Large Model in the legal field.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">At the same time, in the second half of 2024, specific sub-scenario application products (such as customer service) built upon large models achieved rapid growth, increasing from 13 to 89, and in the first half of 2025, there were 127 Agent products registered, exceeding the data from the entire previous year.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">These data reflect the rapid growth of market demand and demonstrate the trend of AI transitioning from \"breadth\" to \"depth,\" indicating that the market is shifting from \"technological showpieces\" to \"scenario deepening.\"\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cstrong class=\"ql-lineHeight-1-75\" style=\"font-size: 18px; color: rgb(255, 153, 0);\">III. Scenario Application Map: Education and Healthcare as Twin Drivers\u003C\u002Fstrong>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">According to IT Juzi statistics, the main application scenarios of domestic AIGC large models include education, healthcare, finance, customer service, industry, and many other fields, covering aspects such as office work, shopping, entertainment, and learning, showing the penetration of AI into the real economy.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cimg alt=\"undefined\" src=\"https:\u002F\u002Fimage.51xinwei.com\u002F2025\u002F09\u002F5809bf8eca324c9fb2da1602ddafea80\u002F749.jpg\" width=\"undefined\" height=\"undefined\" style=\"display: block; margin: auto;\">\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">Among them, there are 39 educational AIGC large models and related applications, ranking first in total, far ahead of other scenarios. Representative examples include NetEase Youdao's \"Ziyue\" large model, Good Future's MathGPT large model, and Yuanli Technology's \"Kanyun\" large model. AI has shone brightly in specific educational scenarios such as personalized teaching and oral practice, responding to the \"double reduction\" policy and online education demands.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">There are 21 medical large models and related AIGC application products, ranking second in total, including Xiaohuo Health's Xiaohuo AI doctor and Haoxinqing Health Industry's \"Haoxinqing AI companion,\" focusing on diagnostic assistance, patient consultation, and psychological healing.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">Companies like Industrial and Commercial Bank of China have also developed AIGC products focused on customer service scenarios using large model capabilities, totaling 16, used to improve the efficiency of intelligent customer service inquiries, commonly found in e-commerce and financial sectors.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">Representative financial large models include Ant Financial's AntGLM large model and Duxiaoman's Xuan Yu large model, as well as Tonghuashun's Wenchai HithinkGPT large model, applied in risk control and financial consulting.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">Based on large model capabilities, there are also many AIGC application services for the cultural tourism industry, such as Mafengwo's \"AI Itinerary Assistant\" and \"Didi Travel Assistant,\" optimizing travel planning and travel services. In addition, Ctrip has released the \"Dao Wen\" vertical cultural tourism large model, and Tongcheng Travel has released the \"Chengxin\" large model.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">The media industry also has Agent smart products, such as Caishulian's Xiaocaishen, used for AI content generation and information interpretation.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">These scenarios reflect the \"pragmatic\" orientation of China's AI. In the future, as multimodal AI (such as text-to-image, voice interaction, video generation) matures, cross-scenario integration will become a trend, driving AI from single applications to an ecosystem.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cimg alt=\"undefined\" src=\"https:\u002F\u002Fimage.51xinwei.com\u002F2025\u002F09\u002F8e290550b0b84e269ec5f60abd66eafa\u002F751.jpg\" width=\"undefined\" height=\"undefined\" style=\"display: block; margin: auto;\">\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cstrong style=\"font-size: 18px; color: rgb(255, 153, 0);\">IV. Regional Distribution: Beijing Leads Entirely, Shanghai and Guangdong Follow Closely\u003C\u002Fstrong>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">The acceleration of AI filings in China is closely related to regulatory policies and the entrepreneurial and innovative environment in various regions.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">2023, as the \"first year\" of generative AI large models, saw most filings concentrated in first-tier cities such as Beijing and Shanghai;\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">From 2024 to 2025, it spread nationwide, with more AIGC products derived from large models emerging in provinces such as Guangdong, Jiangsu, Zhejiang, and Sichuan.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cimg alt=\"undefined\" src=\"https:\u002F\u002Fimage.51xinwei.com\u002F2025\u002F09\u002F56e2ad409dfa48b68c47acd9605435c5\u002F755.jpg\" width=\"undefined\" height=\"undefined\" style=\"display: block; margin: auto;\">\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">Looking specifically, Beijing currently has the most large models and AIGC products, accounting for about 31% of the total, among which there are 73 registered Agent products, and 132 large models have been filed, mainly general large models, such as Baidu's \"Wenxin Yinyi,\"出门问问's \"Sequence Monkeys\" large model, Meituan's \"Tonghui\" large model, etc.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">This fully demonstrates that Beijing companies have leading AI large model R&D capabilities, with technological foresight at the forefront, mainly due to the capital's scientific research resources (such as the Chinese Academy of Sciences) and the concentration of internet giants.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">Shanghai follows closely with a 26% share, among which there are 95 registered Agent products and 82 large models that have been filed, such as SenseTime's \"Ririxin\" large model, the Shusheng Puyu large model released by the Shanghai Artificial Intelligence Laboratory, and the Jie Step Star's Jie step model. This indicates that the large model R&D strength of Shanghai companies should not be underestimated, but their response speed to the commercialization of AIGC products is stronger, with more commercial thinking.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">Guangdong accounts for 11%, among which there are 8 registered Agent products and 66 large models that have been filed, excluding Tencent's Hunyuan Assistant large model and Yuanxiang XVERSE large model, the rest are mainly released by hardware manufacturers and electronics companies, such as vivo's \"Lanxin\" large model, OPPO's AndesGPT-LVM, Honor's Magic large model, ZTE Communications' Xingyun large model, and TCL large model assistant, etc.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">Other provinces such as Jiangsu, Zhejiang, Guizhou, etc., also have registered AIGC products, reflecting balanced regional development.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cstrong class=\"ql-lineHeight-1-75\" style=\"font-size: 18px; color: rgb(255, 153, 0);\">V. R&D Ecosystem: Multilayered Innovation\u003C\u002Fstrong>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">IT Juzi classified and counted the R&D entities of large models, with the following results:\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cimg alt=\"undefined\" src=\"https:\u002F\u002Fimage.51xinwei.com\u002F2025\u002F09\u002Fe2c59d2f45a54f7999f3b3643c81bd96\u002F760.jpg\" width=\"undefined\" height=\"undefined\" style=\"display: block; margin: auto;\">\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">Nearly 100 listed companies are participants in large models, such as Baidu's Wenxin Yinyi, Alibaba's Tongyi, Tencent's Hunyuan large model, and Keda Xunfei, Kunlun Wanwei, etc. They dominate the development of general large models with their financial and technological advantages.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">There are 47 AIGC large models independently developed by domestic unicorn companies, such as Zhigpu AI's several large models - Zhigpu Qingyan, Qingying AI, Zhigpu End-side, Douyin's Yunque large model, Baichuan Intelligent's Baiying large model, MiniMax's abab model, Yueli Dark Face's Moonshot large model, etc., mostly high-valued companies that have not yet gone public.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">Other startups besides unicorns also participated in the development of large models or Agent products, demonstrating innovation vitality.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">IT Juzi noticed that although private tech companies are the main participants in AI large models, other organizational entities including central state-owned enterprises and state-owned enterprises (China Mobile's Jiutian large model, PetroChina's Kunlun large model), universities and research institutions (the Zidongtaiyu large model from the Institute of Automation, Chinese Academy of Sciences), and hospitals have also not missed this competition.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">According to statistics, the participating universities now include: the Tian Gong Kai Wu large model released by the Artificial Intelligence Research Institute of Harbin Institute of Technology, the CivilGPT released by Tongji University, and the Agent intelligent product \"Xiaohong AI Patient Assistant\" released by the Affiliated Obstetrics and Gynecology Hospital of Fudan University, the Agent product \"Ruijin Hospital Medical Intelligent Q&A\" released by the Affiliated Ruijin Hospital of Shanghai Jiao Tong University School of Medicine, and the Agent product \"Guling\" released by Nanjing Drum Tower Hospital, etc.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">This structure highlights the \"multilayered\" ecosystem of China's AIGC: listed companies provide stability, startup companies inject vitality, and research institutions drive fundamental innovation.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">In summary, China's generative AI is entering its \"golden age\": the total number of filings and registrations has soared from 61 in 2023 to a cumulative 672 by the first half of 2025, representing a tenfold increase, reflecting the effectiveness of national strategies (such as the \"New Generation Artificial Intelligence Development Plan\"). The rise of vertical models marks the transformation of AI from \"general tools\" to \"industry empowerment,\" especially in application areas such as education and healthcare, where AIGC products are developing rapidly, reflecting the practical value orientation of AI.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cspan style=\"color: rgb(187, 187, 187);\">[News Source] IT Juzi \u003C\u002Fspan>\u003Cspan style=\"color: rgb(187, 187, 187); background-color: rgb(56, 56, 56);\">Author | Wu Meimei  \u003C\u002Fspan>\u003Ca href=\"https:\u002F\u002Fmp.weixin.qq.com\u002Fs?__biz=MjM5MjQ2NzA2Mg==&mid=2649624127&idx=1&sn=179661e187cf3f2ab004fdb1da39e9be&poc_token=HHbaw2ijGb0jU2lDAxKWi8UP_kaN_URBWsKQbYcT\" rel=\"noopener noreferrer\" target=\"_blank\" style=\"color: rgb(187, 187, 187);\">http:\u002F\u002Fi9n.cn\u002FcvDgK\u003C\u002Fa>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cspan style=\"color: rgb(187, 187, 187);\">（This article is reprinted by this website to provide readers with more information and news. The content involved does not constitute investment or consumption advice. If there are any questions about the facts of the article, please verify with the relevant parties. The views expressed in the article are not the views of this website, and are for reference only.)\u003C\u002Fspan>\u003C\u002Fp>","https:\u002F\u002Fxinwei-dev-test.oss-cn-shenzhen.aliyuncs.com\u002Fintelligent\u002Faudio%3A14fb4f7f-45a0-4af8-9cdb-441d0a805940%3A0.wav?Expires=1774838472&OSSAccessKeyId=LTAI5tNvY2RkKjZw4LLWsrPK&Signature=K0sH8Ds1o5aKprUX76KzTw%2BJNCc%3D","14fb4f7f-45a0-4af8-9cdb-441d0a805940",18459918]