Core Insights - The number of candidates registered for the national college entrance examination (Gaokao) in 2025 has exceeded 13.35 million, marking a historical high, which presents unprecedented challenges for students and parents [1] - The implementation of new Gaokao reforms across 29 provinces has shifted the admissions model towards a more complex strategy involving "institutional major groups" or "major + institution," making the application process more about information and strategy rather than just scores [1] - The rapid development of technology and industry changes significantly influence students' choice of majors and future employment directions, highlighting the need for effective decision-making tools [1] Group 1: Volunteer Application Tools - The "Sunshine Volunteer" application, backed by official data from the Ministry of Education and universities, offers a comprehensive and practical interface for students to navigate the application process [2][3] - Users can receive personalized recommendations based on key metrics such as exam scores and rankings through features like "intelligent screening" and "major inclination assessment" [2][3] - Despite its advantages, users have reported issues with the system's recommendation accuracy and adaptability to new Gaokao rules, indicating areas for improvement [4] Group 2: High School Volunteer Tools - "Gaokao Paper Kite" is a lightweight application that allows users to search for information easily, providing a conversational interface for inquiries [5][8] - The application emphasizes a free and convenient approach, catering to users who need quick answers and initial guidance [9] - "Quark" offers a detailed input process for user preferences, enabling more precise recommendations based on individual goals and strategies [11][14] - "Quark" has a large user base, having served over 120 million candidates and parents, indicating its popularity and relevance in the market [24] Group 3: Paid Applications - "Youzhiyuan" provides advanced features such as AI-driven recommendations and risk analysis for college applications, enhancing user experience [35][39] - The application introduces a unique "major group purity" metric to help users minimize the risk of being assigned to less desirable majors [39] - "Reliable AI" claims a significant improvement in prediction accuracy for college admissions, utilizing extensive data analysis to enhance user decision-making [41][44] Group 4: Market Trends and User Needs - The increasing number of candidates and the complexity of new Gaokao reforms necessitate the continuous evolution of volunteer application tools to meet diverse user needs [24][47] - The tools range from free applications to paid services, catering to various levels of user requirements, from basic information to in-depth personalized guidance [47] - Ultimately, the effectiveness of these tools relies on the candidates' ability to integrate the information provided with their personal aspirations and circumstances [47]

以下文章来源于富国基金 ，作者填志愿 富国基金 . 因此，最近各个社交平台都上演着"志愿填报血泪史"，小编看到不少家长和考生线上发帖，有纠 结"专业如何选"、"热门专业是哪些"，也有焦虑"冲一冲会不会滑档"、"985有点坑的专业VS双非王 牌专业如何选"……刷着这些充满困惑的讨论，小编深有感触：无论是选学校选专业，还是选基金， 一定程度上可以说都是对未来的投资，细细品味，其实这志愿填报和买基金的弯弯绕绕颇有异曲同 工之处~ ★ 第一波 ★ 难选程度——眼花缭乱？ 高考选择院校、专业之难，就如同选基一样，通过下面这组数据可见数量之繁多。 志愿填报： 根据教育部2025年最新发布的普通高等学校本科专业目录，共包含93个专业类，845种专业， 其中今年新设了29种新专业，分别涉及区域国别学、碳中和科学与工程、海洋科学与技术、 航空运动、智能分子工程、人工智能教育等方向。同时，截至2024年6月20日，全国高等学校 共计3117所，其中普通高等学校2868所。可见院校林立，选择维度之复杂。 选基： 截至2025年6月25日，数据显示，共有12869只公募基金（仅含主代码），其中非货12499 只，基金经理共有4043 ...

炒作高分考生"树欲静而风不止"，归根结底，还在于单一的教育评价体系尚未得到根本扭转。当"行行 出状元"真正成为共识，当评价制度为艺术、体育、创新思维、实践能力等特质真正留出空间，围绕高 分考生的炒作才可能降温。 其实，推行十余年的高考改革，核心目标正是分类考试、综合评价、多元录取，打破"唯分数论"。如 今，在高考分数之外，各地建立了省级统一的综合素质评价信息管理系统，评价结果已成为"强基计 划"等招生方式的重要参考。 要继续深化教育评价改革，引导社会树立正确的教育观和成才观。教育部门要真正将升学率与学校工程 项目、经费分配、评优评先等脱钩，学校要将育人重心转向立德树人，促进学生全面发展。毕竟，教育 的意义不在于缔造传奇，不在于培养少数"高分者"，而在于让每个普通人都有人生出彩的机会。 我们要为高分考生鼓掌，也要为所有奋力拼搏的学子鼓掌。愿你们在一个终身学习的社会里，继续前 进，探索不同的可能性，书写属于自己的故事。 每到高考成绩放榜日，"高分考生"便成为关注焦点。尽管教育部连年重申，不得以任何形式宣传炒 作"高考状元""升学率""录取率""高分考生"，多地也用"分数屏蔽"的方法试图阻断对状元的"围追堵 截"， ...

为更好地服务考生和家长，助力高校招生宣传工作，今年全省7个市开展陕西高校校园开放日系列 公益活动。 在西安电子科技大学校园开放日现场，来自省内外数十家高校的招办工作人员为考生和家长答疑解 惑。 "2025年学校再次增加招生110人，新增的招生计划主要投放在通信、电子工程、集成电路、人工智 能、网络安全等国家急需和报考热门专业领域。"西安电子科技大学招办负责人介绍，学校充分发挥电 子与信息专业特色鲜明优势，聚焦电子信息不同领域设置招生专业类，比如电子信息类分为通信、电子 工程、集成电路等专业类，计算机类分为计算机科学、网络安全、人工智能等专业类。学校还将报考热 度高的计算机科学与技术（国家拔尖基地班）专业进行单独招生，同时新增了数据科学与大数据技术中 外合作办学项目。（记者：吕扬） 6月26日，省教育厅及各高校通过各种形式，积极为考生提供填报志愿服务。 "考虑到今年是新高考第一年，家长和考生对相关政策不是很了解，学校为考生着想，将相似、相 近专业分为一组，每个组专业不超过6个，且这些专业热度和往年招生位次接近。"付刚说，学校响应国 家需求，今年网络安全这个校企联办的热门专业将招收200人，在陕投放120个名额， ...

Core Viewpoint - The rise of AI-powered college application tools is transforming the way students approach the college admission process, with major tech companies like Baidu, Alibaba, and Tencent launching free applications to assist students in filling out their college applications [1][2]. Group 1: AI College Application Tools - Major internet companies have introduced AI college application tools that provide personalized recommendations based on students' input, such as province, subject, and scores, marking a shift from traditional methods [2][3]. - The tools utilize large model technology to analyze admission data and offer tailored suggestions, enhancing the user experience compared to previous years [2][3]. - Some platforms have integrated personality tests to suggest suitable majors, although these recommendations are still basic and rely on simple matching [2][3]. Group 2: Business Strategy Behind Free Services - Despite claiming not to profit from these services, major companies embed these tools within their browsers to capture user traffic, indicating a strategic move to dominate the market [3][4]. - Data from these companies shows significant user engagement, with Baidu reporting over 95% of families using its service during the exam period, and QQ Browser claiming to have served 400 million users [3][4]. - The ultimate goal for these companies is to retain users and create future monetization opportunities through AI models or search services [3][4]. Group 3: Effectiveness and Concerns - There are notable discrepancies in the recommendations provided by different platforms, raising questions about the reliability of the algorithms used [4][5]. - The lack of transparency regarding the recommendation algorithms has led to confusion among users, as they receive conflicting advice [4][5]. - Experts suggest that while AI tools can provide valuable data analysis, they should not replace personal decision-making, and students should consider multiple factors when making their final choices [5].

以下文章来源于财经图解 ，作者乘桴于海 山东考生的情况则相对较好，2023年山东本地高校本科招生32.62万人，该年山东高考人数98万人，同 样假设的情况下本科招生比例29.19%。 财经图解 . 此外，四川、湖南、安徽、江西等高考人数位居前列的省份，本科竞争烈度也相对较大。同样假设的情 况下，2023年本地高校本科招生数与该年高考人数的比例均在1/3左右。 财经大事早知道，关注东方财富股票！ 截至6月25日，全国31个省份均已公布高考分数线，全国各地考生和家长们正紧锣密鼓地进入志愿填报 时间，高考相关话题再度成为舆论焦点。 教育部数据显示，2025年全国高考报名人数1335万人，比2024年减少7万人，这是8年来全国高考报名人 数首次下降。而2024年，普通本科招生489.97万人，本科录取率为36.5%。 考生竞争烈度方面，河南、山东今年高考人数均超百万。河南考生人数继续位列全国第一，有136万人 参加高考；山东考生101万人，为史上首次突破百万。 庞大的考生数也由此引发了高强度的竞争。教育部数据显示，2023年，河南本地高校本科招生数为32.6 万人，而该年河南高考人数为131万人。假设本地考生全部报告 ...

随着全国高考成绩陆续公布，志愿填报成为全社会关注的焦点。 6月26日，张朝阳和张雪峰在一场对谈中深入探讨了高考志愿选择话题。张雪峰强调，高考志愿选择没 有绝对的最好，关键在于是否适合个人情况。他指出，考生应在有限的时间内提升信息搜索和分析的效 率，打破信息差，找到最适合自己的方案。 兴趣与就业如何权衡？追求星辰大海也要脚踏实地 张朝阳则以教育本质为切入点，他认为大学教育是一种基础教育，未来的职业成功与选择的专业并无直 接关联，可以在考虑自己兴趣的同时，兼顾未来就业方向。他提出"50%考虑就业，50%考虑求知和兴 趣"的平衡法则：兼顾兴趣与就业方向的折中选择并非中庸，而是一种平衡与双赢的选择。他以自身经 历为例，提到自身是麻省理工学院物理博士毕业两年后转向互联网，但基础学科培养的逻辑思维始终受 用，强调"不必因专业限制而否定未来可能性"。 善用人工智能作为辅助工具，结合实际情况推出最佳报考方案 当前，人工智能技术蓬勃发展，相关技术也成为了高考志愿填报的辅助工具。在节目现场，张朝阳与张 雪峰一起体验了一款AI志愿报告功能。考生只需输入省份、分数、选科、兴趣偏好等信息，就能快速 生成包含院校优先、就业优先、地域优先 ...

编辑：杨文、+0 机器之心报道 测试全程未做任何 prompting engineering，所有输入均为高考原题，其中 DeepSeek R1 输入为题目文本，其 余模型则是题目文本和题目截图。在总分计算上，采用 3（语数外）+3（理综 / 文综）的形式对 5 个模型进 行排名。 从最终成绩单来看，这 5 家大模型的文科成绩均超 620 分， 如果按照山东高考的赋分制，豆包的 683 分可 以冲刺清华、北大 ；在理科方面，各大模型之间的分数差距则较为明显， Gemimi 和豆包已达到保底重点 985 的水准，而 Claude 4 和 o3 还不及 600 分。 去年高考全科测评中，大模型们还只能勉强踩到一本线，面对复杂的数学、物理题目时，虽然能产出答 案，但思路浅显、推理链条不够严密，常常给人一种「全靠蒙」的感觉。然而短短一年过去，技术更新带 来了质的飞跃，大模型展现出越来越强的逻辑推理和解决深度问题的能力。 语数英区分度较小，理科总分不及文科 在语、数、外等基础学科上，参评模型整体表现优异，均已达到顶尖考生水平，彼此间的区分度相对较 小。不过，o3 模型因作文跑题导致语文单科得分偏低，拖累了其总分。 今年 ...

这几天，2025年全国高考考生分数陆续公布，各省市高考批次线也相应公布，高考和填报志愿就像一场 接力赛，选择学校和专业这一棒能否接好，关系到考生未来的成长赛道和人生规划。 接下来，deepseek提示需要参考往年的数据，重点参考2024年、2023年甚至2022年北京考生报考各高校 金融类专业（或经济学大类）的最低录取位次和平均录取位次。而且还提示，目标院校的招生官网、北 京市教育考试院发布的往年录取数据是权威来源。具体到报什么专业，deepseek给出了几种分类，金融 专业可以以不同的名称招生，如金融学、经济学类、投资学等。 以上基本信息确认后，deepseek认为，考生这一分数比较微妙，"冲刺顶尖金融名校仍有难度，但放弃 又不甘心"，并且需要帮用户在"冲名校边缘专业"和"保金融核心专业"之间找平衡点。综合以上情况， deepseek的核心策略确定为"重点保专业实力和北京区位"，即包含冲刺策略（需关注大小年）、稳健匹 配（位次高度契合）、高性价比保底（锁定金融专业）。 具体来看，在冲刺策略中，deepseek推荐中国政法大学的金融工程专业、西南财经大学的金融学（证券 与期货方向），其中中国政法大学的优势在于 ...

Core Insights - ByteDance's Seed team released the results of the 2025 college entrance examination simulation, showcasing the performance of their model, Doubao Seed 1.6-Thinking, which achieved scores of 683 in liberal arts and 648 in sciences, ranking it highest in liberal arts among five top models tested [1] - The results indicate that Doubao excelled in subjects such as Chinese, English, Physics, History, Geography, and Politics, with a mathematics score exceeding 140. The DeepSeek R1 model achieved the highest score in mathematics at 145, while Gemini 2.5 Pro scored 82 in Chemistry and tied with OpenAI o3 for the highest score in Biology at 77 [1] - The Seed team noted that the initial test used low-resolution exam papers, which affected performance in Chemistry and Biology, but after obtaining higher quality images, Doubao's scores improved by nearly 30 points in these subjects, raising the total science score to 676, demonstrating the potential of multimodal reasoning [1] Subject Performance Summary - The Shandong province's college entrance examination follows a "3+3" format, with core subjects being Chinese, Mathematics, and English, and three elective subjects scored on a grade conversion basis. The highest possible score for Doubao's combination could reach around 690, potentially placing it in the top 80 rankings for admission to prestigious universities like Tsinghua and Peking University [2] - Performance breakdown of various models in the examination: - Total score: Doubao Seed 1.6-Thinking scored 648, while Gemini 2.5 Pro scored 655, DeepSeek R1 scored 615, Claude scored 268, and OpenAI o3 scored 579 [2] - Subject scores: - Chinese: Doubao 128, Gemini 126, DeepSeek 118, Claude 117, OpenAI 135 - Mathematics: Doubao 741, DeepSeek 145, Claude 128, OpenAI 136 - English: Doubao 144, Gemini 141, DeepSeek 143, Claude 142, OpenAI 140 - Physics: Doubao 90, Gemini 86, DeepSeek 62, Claude 74, OpenAI 78 - Chemistry: Doubao 78, Gemini 82, DeepSeek 72, Claude 61, OpenAI 71 - Biology: Doubao 76, Gemini 77, DeepSeek 75, Claude 76, OpenAI 77 - Geography: Doubao 94, Gemini 78, DeepSeek 19, Claude 92, OpenAI 90 - History: Doubao 92, Gemini 84, DeepSeek 67, Claude 78, OpenAI 84 - Politics: Doubao 84, Gemini 82, DeepSeek 79, Claude 76, OpenAI 80 [2]