证券日报网讯12月24日，华宇软件（300271）在互动平台回答投资者提问时表示，截至2025年11月28 日，公司股东总户数(未合并融资融券信用账户)为46085户。截至12月底的股东人数，预计可以在1月5 日(周一)通过互动易提问答复。 ...

Group 1 - The core point of the article is that Jin Modern (SZ 300830) held its 13th meeting of the fourth board of directors on December 24, 2025, to discuss the proposal for using part of the idle raised funds for cash management [1] - For the first half of 2025, Jin Modern's revenue composition is as follows: electricity accounts for 87.79%, other sectors for 9.68%, and aerospace for 2.53% [1] - As of the time of reporting, Jin Modern has a market capitalization of 4.5 billion yuan [1]

证券日报网讯 12月24日，旗天科技在互动平台回答投资者提问时表示，公司定增事项进展情况请关注 深交所发行上市审核信息公开网站的相关信息和公司在指定信息披露媒体上刊登的公告。 （文章来源：证券日报） ...

Group 1 - The core point of the article is that the company, Gao Weida, reported that as of December 10, 2025, it had a total of 50,111 shareholders [2]

证券日报网讯 12月24日，宝信软件在互动平台回答投资者提问时表示，公司股价波动受到宏观经济、 行业政策、市场情绪及资金供求等多重复杂因素的综合影响。公司始终秉持规范运作原则，严格履行信 息披露义务。公司将始终坚持聚焦主业，通过扎实做好生产经营、持续提升核心竞争力和内在价值来积 极应对，力争实现稳健发展，以切实的经营业绩回报广大投资者的信任。 （文章来源：证券日报） ...

Core Viewpoint - The company has transitioned from collaborative development with the Chinese Academy of Sciences to primarily independent research and development of its Morning Star large model, while also incorporating some open-source models based on user needs [2] Group 1 - The Morning Star large model was initially developed in collaboration with the Chinese Academy of Sciences [2] - The company is now focusing on independent development of the Morning Star large model [2] - The company is utilizing some open-source large models according to user requirements [2]

每经记者｜张宏 朱恒 每经编辑｜魏官红 "监控神器"被公开售卖 销售：员工是察觉不到的 "如何神不知鬼不觉地对员工电脑屏幕进行监管？其实非常简单，只需要安装这款软件，就可以监管到公司所有电脑，实时查看到员工电脑屏幕。员工从 安装开始全程无感知、无弹窗⋯⋯" 小红书上，有销售人员发布多个宣传视频，展示安装在员工电脑里的"电子眼"的强大监控功能。 《每日经济新闻》记者（以下简称每经记者）发现，一款面向公司用户、单套价格或不足300元的监控软件正在被公开售卖。 "笔记本、台式机都能用⋯⋯可以等员工下班了再安装，安装好之后把安装包删除⋯⋯被监控过程中，员工那边是察觉不到的，咱们有这个优点。"对于这 款软件，河北安固云软件有限公司（以下简称安固）的销售人员介绍称。 0:00 / 2:52 每经记者：受控端可以看到安装文件吗？ 安固销售：看不到，到时候会隐藏起来。 每经记者：杀毒软件查得出来吗？ 安固销售：我们老板已经和这些厂商协调过，都是兼容的。这不是病毒，所以（杀毒软件）是检测不到的。 "可以同步监管到员工微信和QQ""无感知""无弹窗"⋯⋯"监控神器"真的能"完美隐身"？每经记者展开深入调查。 记者注意到，通过安固 ...

（文章来源：每日经济新闻） 每经AI快讯，有投资者在投资者互动平台提问：公司有海底数据中心的项目建设计划吗？ 宝信软件（600845.SH）12月24日在投资者互动平台表示，公司目前无此项计划。 ...

声明：市场有风险，投资需谨慎。本文为AI基于第三方数据生成，仅供参考，不构成个人投资建议。 来源：市场资讯 鼎捷数智公告称，根据中国证券监督管理委员会《关于同意鼎捷数智股份有限公司向不特定对象发行可 转换公司债券注册的批复》（证监许可〔2025〕2297号），公司向不特定对象发行可转换公司债券 8,276,642张，每张面值人民币100.00元，发行总额人民币827,664,200.00元，扣除承销保荐等费用后实际 募集资金净额为人民币815,690,054.97元，上述资金已于2025年12月19日划至公司指定账户。为规范募 集资金的存放、使用和管理，公司及全资子公司湖州鼎捷软件有限公司与兴业银行股份有限公司上海黄 浦支行、保荐机构兴业证券股份有限公司签署了《募集资金三方监管协议》，并开立三个募集资金专项 账户，分别用于鼎捷数智化生态赋能平台项目、补充流动资金项目等。 ...

Investment Rating - The investment rating for the industry is "Positive" and is maintained [7] Core Insights - On December 19, OpenAI released GPT-5.2-Codex, which is built on the general model GPT-5.2 and optimized for "Agentic Coding" scenarios, primarily targeting complex software engineering tasks. The new model has systematic improvements in long-range task execution, large-scale code changes, Windows native environment support, and network security capabilities compared to previous versions. The commercialization of large models is expected to accelerate, with current costs being a core factor limiting token consumption. Attention should be paid to the cost reduction effects of large models [2][5][9] Summary by Sections Event Description - OpenAI's GPT-5.2-Codex is specifically optimized for complex software engineering tasks, showing systematic improvements in long-range task execution, large-scale code changes, and Windows native environment support [5][6] Performance Insights - GPT-5.2-Codex introduces a native context compression mechanism, enhancing understanding and utilization efficiency for long contexts, making it more stable for long-term coding tasks. The model shows improved performance in code refactoring and migration scenarios, with enhanced reliability and consistency [9] - In terms of network security, GPT-5.2-Codex is the strongest Codex model from OpenAI, demonstrating excellent performance in various security tests, although it has not yet reached the highest risk threshold. The model's capabilities can meet legitimate protection needs while avoiding excessive risks [9] - The model's performance metrics include an accuracy of 56.4% in SWE-BenchPro and 64.0% in Terminal-Bench2.0, indicating a significant improvement in reasoning token efficiency, making it suitable for long-term coding tasks with broad commercialization prospects [9] Future Outlook - The report emphasizes the importance of monitoring future model releases and the need to focus on cost reduction effects for commercialization. The domestic AI industry chain is viewed positively, with continued recommendations for shovel stocks and major players with significant positioning advantages [9]