Core Viewpoint - The company has made significant investments in optimizing the training and inference cost efficiency of its large models under limited computing resources, achieving substantial improvements in performance metrics compared to industry standards [1] Group 1: Technological Advancements - Since May 2023, the company has collaborated with Huawei to overcome various technical challenges, including high-speed interconnection, hidden computation communication, and optimization of training and inference efficiency [1] - The training efficiency of general large models and deep inference models has improved from an initial 30%-50% to over 85%-95% when benchmarked against NVIDIA's A100 [1] Group 2: Breakthroughs in Domestic Computing Power - In 2025, the company achieved significant breakthroughs in two areas: enhancing the training efficiency of long-thought chain reinforcement learning from 30% to over 84% against the A800 benchmark, and improving the full-link training efficiency of MoE models from 30% to 93% [1] - These advancements represent a major leap from 0 to 1 in the domestic computing power sector, indicating a strong potential for further cost reductions in training [1]

Core Viewpoint - The company is enhancing its AI services through the upgraded iFlytek Robot Super Brain platform, which focuses on multi-modal perception and interaction capabilities, aiming to assist robot manufacturers in rapid integration and development [1] Group 1: Technology and Features - The iFlytek Robot Super Brain platform is built on the iFlytek Super Brain 2030 technology foundation, providing core AI services to robot manufacturers [1] - The platform's upgrade, scheduled for October 28, 2025, addresses real-world interaction challenges and industry development needs, featuring breakthroughs in multi-person multi-modal noise reduction technology, high emotional intelligence dialogue experience, precise response to continuous commands, and a comprehensive solution for embodied intelligence [1] - The platform allows for low-threshold access through an integrated hardware and software approach, facilitating quick adoption by manufacturers [1] Group 2: Industry Collaboration - The company has established partnerships with over 500 intelligent robot manufacturers, including leading firms such as Yushu, Zhiyuan, Galaxy General, and Songyan Power [1] - The collaboration spans various types of robots, including humanoid robots, quadruped robots, and public service robots, indicating a broad industry reach [1]

（编辑 丛可心） 证券日报网讯 1月6日，科大讯飞在互动平台回答投资者提问时表示，在人工智能产业爆发的关键窗口 期和重大机遇期，科大讯飞会保持对科技创新的定力和在关键战略方向上的投入，通过保持单点核心技 术的持续创新和领先，不断拓展科技树的广度和深度进行系统性创新，深入教育、医疗、城市等国家重 大民生需求，汽车、金融、消费品等国民经济重要场景，持续落地规模化应用推广人工智能产品和方 案，提高可持续发展和长期盈利能力。具体业绩情况敬请关注公司定期报告。 ...

证券日报网讯1月6日，科大讯飞（002230）在互动平台回答投资者提问时表示，2025年12月16日，科大 讯飞与中国进出口银行签署战略合作协议，未来双方将在技术攻关、海外布局、新型政策性金融工具运 用等方面进一步深化合作，实现互利共赢，共同助力推进落实《全球发展倡议数字普惠行动倡议》和 《人工智能全球治理行动计划》；12月19日，科大讯飞与中国银联签署创新业务合作协议，未来双方将 围绕"人工智能+金融支付"，重点推进国家人工智能应用中试基地共建、金融支付垂域大模型研发、场 景示范应用开发等联合创新，为支付产业发展注入新动力。 ...

证券日报网讯1月6日，科大讯飞（002230）在互动平台回答投资者提问时表示，科大讯飞多年来一以贯 之的脚踏实地推进人工智能技术与产业进步，未来会继续保持"1+N"的战略投入定力即"1个底座大模型 +N个行业大模型"的整体布局，底座大模型持续对标国内外最高水平，同时围绕赋能教育、医疗等关键 领域构建N个重点行业大模型，积极追求用最先进和安全的人工智能赋能国计民生和国家战略领域。 ...

本报讯 （记者袁传玺）天眼查工商信息显示，近日，讯飞和光（郑州）科技有限公司成立，注册资本50万元，经营范围含 信息技术咨询服务、软件开发、农产品智能物流装备销售、信息系统集成服务、物联网设备销售、智能机器人的研发、人工智 能应用软件开发、人工智能基础软件开发、人工智能通用应用系统等。股东信息显示，该公司由科大讯飞旗下上海讯飞和光科 技有限公司全资持股。 （编辑 黄力） ...

Group 1 - The company has collected over 1 million hours of data from more than 3,000 households using Gen DAS Gripper and the industry's first large-scale data production line, Gen ADP, accelerating the development of the embodied industry [1] - The overall industry volume and individual skill scale are the largest, with a cumulative duration exceeding 10,000 hours and a total data scale of over 1 million clips, ensuring each skill has the highest data volume in the industry [1] - The data collection focuses on 10 common household scenarios and 30 skills, ensuring quality by avoiding excessive dispersion [1] Group 2 - The data includes multi-modal, high-quality information such as large FOV raw images, trajectories, annotations, and joint actions, with industry-leading accuracy and quality [1] - The first phase of data has been uploaded, covering 12 skills across 4 major scenario tasks, with plans to complete the remaining skills and provide support for data format conversion and usage [2] - The data is available for download at a specified link, and users are encouraged to provide feedback and suggestions for additional scenarios and skills to be digitized [2]

Group 1 - The core point of the news is the establishment of Xunfei and Heguang (Zhengzhou) Technology Co., Ltd., which is fully owned by iFLYTEK Co., Ltd. [1] - The legal representative of the new company is Wang Xin, and its business scope includes sales of IoT devices, research and development of intelligent robots, and development of AI application software [1][2] - The company is registered in Zhengzhou High-tech Industrial Development Zone, Henan Province, with a registered capital of 500,000 yuan [2] Group 2 - The company is classified under the national standard industry of information technology consulting services [2] - The business activities include technology services, technology development, and sales of artificial intelligence hardware and software [2]

Core Insights - Xunfei and Guang (Zhengzhou) Technology Co., Ltd. was established on December 26, 2025, with a registered capital of 500,000 RMB [1][3] - The company is located in Zhengzhou, Henan Province, and is engaged in the sales of IoT devices, research and development of intelligent robots, and development of AI application software and foundational software [1][3] - The company is wholly owned by Shanghai Xunfei and Guang Technology Co., Ltd., which is a subsidiary of iFlytek, with Liu Qingfeng, the founder, chairman, and CEO of iFlytek, as the actual controller [1][3]

Group 1 - As of January 5, the margin trading balance in A-shares reached 25,606.48 billion yuan, an increase of 199.66 billion yuan from the previous trading day, accounting for 2.55% of the A-share circulating market value [1] - The trading volume of margin transactions on the same day was 2,915.25 billion yuan, which is an increase of 730.7 billion yuan from the previous trading day, representing 11.34% of the total A-share trading volume [1] - Among the 31 primary industries in the Shenwan index, 27 industries experienced net financing inflows, with the electronics industry leading with a net inflow of 3.35 billion yuan [1] Group 2 - A total of 63 individual stocks had net financing inflows exceeding 100 million yuan, with Shenghong Technology leading at a net inflow of 694 million yuan [2] - Other notable stocks with significant net financing inflows include Lioo Co., Zhongwei Company, Yanhai Co., Aerospace Power, Xunwei Communication, Shannon Chip, Keda Xunfei, Western Materials, and Bolite [2] - Guosen Securities research report indicates that the current period is an earnings window, and certain stocks benefiting from high growth in overseas AI computing power are expected to see performance forecasts catalyze in January [2]