文档内容
主题 03 DeepSeek 的发展与应用
主题简析
在初中英语阅读理解中，有关 DeepSeek的发展与应用主题 通常涵盖如下内容：(1)发展历程与技术创新 ：
文章常围绕DeepSeek的成立背景（如2023年成立时间 ）、核心技术（自然语言处理、数据分析等 ）及突
破性成果（如多层动态防御系统对抗网络攻击的案例 ）展开，强调其“引领智能解决方案未来”的科技属
性 。(2)教育领域应用 ：阅读材料多聚焦DeepSeek在教育场景中的实际功能，例如通过AI分析学习习惯提
供个性化学习方案 、实时反馈纠错机制提升学习效率 ，以及辅助教师进行教学管理（如实时跟踪学生进
度 ）等，体现技术与教育的深度融合 。(3)社会价值与挑战 ：文本常结合青少年成长与社会责任主题，探讨
DeepSeek对青少年学习方式的影响（如提升阅读敏感度 ），或引发对科技伦理的思考（如网络安全问题 ），
引导学生辩证看待技术发展 。
有关DeepSeek的阅读材料兼具科技前沿性与教育实用性，通过真实案例与跨学科视角培养学生综合
语言能力与社会认知 。此类阅读题常设置 细节理解 （如技术应用的具体步骤 ）、 词义推断 （如专业术语
“dynamicdefensematrix”）及 观点态度分析 （如对AI利弊的讨论 ），旨在考查信息提取与批判性思维 。
词句积累
主题词汇示例
1.关于DeepSeek的发展
development发展
state-of-the-arttechnology尖端技术，用来形容DeepSeek的技术水平非常先进
continuousimprovement持续改进，表示DeepSeek在不断进步和优化
integrate集成，如“integrateDeepSeekintoeducationalplatforms”（将DeepSeek集成到教育平台中）
2.关于DeepSeek的应用
application应用
AIassistant人工智能助手，DeepSeek可以作为学习助手帮助学生
personalizedlearning个性化学习，DeepSeek能根据学生的学习需求提供定制化的学习内容
languagelearning语言学习，DeepSeek可以辅助学生学习英语等语言
knowledgeacquisition知识获取，学生通过DeepSeek可以更高效地获取知识
educationaltool教育工具，DeepSeek被视为一种创新的教育工具
3.具体应用场景
summarizenews总结新闻，DeepSeek可以帮助学生快速了解新闻要点
generatereports生成报告，学生可以利用DeepSeek整理学习报告
createmindmaps创建思维导图，DeepSeek可以辅助学生制作学习内容的思维导图
analyzedata/分析数据，DeepSeek能帮助学生分析学习数据，找出学习弱点translatesentences翻译句子，DeepSeek可以提供翻译服务，帮助学生理解外语材料
4.学习技巧
askclearquestions提出清晰的问题，这是使用DeepSeek等AI工具时的关键技巧
usespecificprompts使用具体的提示词，以便DeepSeek能更准确地理解你的需求并提供帮助
拓展阅读
DeepSeek is reshaping AI industry
TherapidriseofDeepSeek,aChineseartificialintelligencecompany,isreshapingtheAIindustryand
promptingmarketplayerstoexplorenewpossibilities,expertssaidatthe2025GlobalDeveloperConference,
whichconcludedSunday.
DeepSeekspecializesinlargelanguagemodels,particularlyincoding-relatedAI,andaimstoprovidehigh-
performancemodelswithefficienttraining.Thecompany,basedinHangzhou,Zhejiangprovince,gainedglobal
attentionlastmonthafterlaunchingitsAIreasoningmodels.Themodelsarefullyopensource,cheapertotrainand
performonparwithleadingglobalcounterparts.
"IhopenotjustChineseAIcompanies,butotherglobalAIcompanieslearnfromwhatDeepSeekisdoing.
Startingbyopen-sourcingyoursoftware…they'vemadeitreallyfinanciallyapproachabletobuildyourown
languagemodels,"saidAndrewAitken,atechnicaloversightcommitteememberoftheLinuxFoundation'sFINOS
Foundation."They'vedonethatreallywell,andtherestoftheworldcanlearnfromthat."
IndustryexpertsattheconferenceexpressedadmirationforDeepSeek-R1,oneofthecompany'slatestAI
developments.
"DeepSeek'sopen-sourceadoptionhassetarolemodelforhowAIcanbenefiteverybody,"saidGuRuiquan,
productdirectoratLanyunTechnology."IthelpscreateanecosystemtobuildbetterAIproductsbecauseeverybody
canstandontheshouldersofgiants."
Guaddedthatascostsdecline,demandforcomputingpowerisrisingacrosstheindustry,creatingnew
opportunitiesforAIdevelopment.
ExpertssaidDeepSeek'sopensourceapproachisdrivingthegrowthofsimilarmodels,acceleratingAI
applicationsacrossvariousindustries."Iamextremelyexcitedabouttheoutlookoflargemodels,whichcannowbedevelopedinacheaper,faster,
betterandmoreefficientway,"saidJiaAnya,productdirectoratSenseTime.
"TherecentbreakthroughsinAIindicatethatintegrationcanhappenmorequickly,reasoningcanbe
acceleratedwithlargemodels,AIcapabilitiescanbefurtheroptimizedandbetterAIapplicationswillbe
developed,"Jiasaid.
ShenHaozhan,analgorithmexpertatOMAILab,saidDeepSeek'ssuccesshighlightshowAImodelscanbe
developedatsignificantlylowercostsandprovidesaroadmapforothercompaniestotraintheirownmodels.
"IampersonallyinspiredbyDeepSeek'sopen-sourcestrategy.Itgivesusnewresearchdirectionsbeyond
conventionalmethodologies,whichisahugecontributiontotheopen-sourcecommunity,"Shensaid.
The2025GlobalDeveloperConference,heldfromFridaythroughSundayinShanghai,wasthemed"AI
ShapingtheWorld,UnlockingInfiniteOpportunities."
Theeventbroughttogetherglobaldeveloperstocollaborate,innovateandexploreAIapplicationswhile
promotingcommercializationefforts.
"ArtificialintelligencewillcontinuetobeoneofShanghai'sleadingindustries,"saidChenJie,vice-mayorof
Shanghai,duringtheconference'sopeningceremonyonSaturday.
ChennotedthatundertheguidanceoftheMinistryofIndustryandInformationTechnology,Shanghaihas
madesignificantstridesintechnologicalinnovation,industry-financecooperationandinternationalpartnerships.
Thecity'sAImarketreachedascaleofmorethan450billionyuan($62billion)lastyear,andithashostedthe
WorldArtificialIntelligenceConferenceforsevenconsecutiveyearsandfiled60largemodels.
"I'mreallyimpressedwithwhatShanghaihasdonearoundAI.LargecorporationsareinvestinginAI,andthe
governmentisprovidingincentives,supportandinvestmenttogrowthedevelopercommunityandfocusonopen
source.Shanghaiisreallypositioningitselfasaleaderinthisspace,"Aitkensaid.
Asopen-sourceAImodelscontinuetoevolve,Shanghaiwillleverageitsstrengthsasamegacityand
accelerateeffortstoestablishitselfasaninternationallyinfluentialAIhub,Chenadded.（来源：China Daily）
参考译文：
DeepSeek正在重塑人工智能行业
专家在周日结束的2025年全球开发者大会上表示，中国人工智能公司DeepSeek的迅速崛起正在重塑
人工智能行业，并促使市场参与者探索新的可能性。
DeepSeek专注于大型语言模型，特别是与编码相关的AI，旨在提供具有高效训练的高性能模型。这
家总部位于浙江省杭州市的公司上个月推出了人工智能推理模型，引起了全球的关注。这些模型完全开
源，训练成本更低，性能与全球领先的同行相当。
Linux基金会FINOS基金会技术监督委员会成员AndrewAitken表示：“我希望不仅是中国的人工智能
公司，其他全球人工智能公司也能从DeepSeek的做法中学习。从开源你的软件开始……他们已经使建立
自己的语言模型在财务上变得非常可行。”。“他们做得很好，世界其他国家可以从中吸取教训。”
与会的行业专家对DeepSeek-R1表示钦佩，这是该公司最新的人工智能发展之一。
蓝云科技产品总监顾瑞泉表示：“DeepSeek的开源采用为人工智能如何造福每个人树立了榜样。”。
“它有助于创建一个生态系统来构建更好的人工智能产品，因为每个人都可以站在巨头的肩膀上。”
顾补充说，随着成本下降，整个行业对计算能力的需求正在上升，为人工智能的发展创造了新的机遇。
专家表示，DeepSeek的开源方法正在推动类似模型的增长，加速各个行业的人工智能应用。
商汤科技产品总监贾安雅表示：“我对大型模型的前景感到非常兴奋，现在可以以更便宜、更快、更
好、更高效的方式开发大型模型。”。
贾说：“最近在人工智能方面的突破表明，集成可以更快地发生，推理可以通过大型模型加速，人工
智能能力可以进一步优化，更好的人工智能应用将得到开发。”。
OM人工智能实验室的算法专家沈浩展表示，DeepSeek的成功突显了如何以更低的成本开发人工智能
模型，并为其他公司训练自己的模型提供了路线图。
沈说：“我个人受到DeepSeek开源战略的启发。它为我们提供了超越传统方法的新研究方向，这对
开源社区做出了巨大贡献。”。
周五至周日在上海举行的2025年全球开发者大会的主题是“人工智能塑造世界，解锁无限机遇”
该活动汇集了全球开发人员，共同合作、创新和探索人工智能应用，同时促进商业化努力。
“人工智能将继续成为上海的主导产业之一，”上海市副市长陈杰在周六的会议开幕式上表示。
陈指出，在工业和信息化部的指导下，上海在技术创新、产融合作和国际伙伴关系方面取得了长足进
步。去年，该市的人工智能市场规模超过4500亿元人民币（620亿美元），连续七年举办世界人工智能大
会，提交了60个大型模型。
艾特肯说：“上海在人工智能方面所做的工作给我留下了深刻的印象。大公司正在投资人工智能，政
府正在提供激励、支持和投资，以发展开发者社区，并专注于开源。上海真的将自己定位为这一领域的领
导者。”。
陈补充道，随着开源人工智能模型的不断发展，上海将利用其作为特大城市的优势，加快建设成为具
有国际影响力的人工智能中心。
综合演练
一、阅读理解
1
January 20, 2025, a Chinese tech company named DeepSeek made a new AI called DeepSeek-R1. This
smart program can solve math problems, write code (代码), and answer questions like top models such as
OpenAI’s GPT-4o, but it costs much less to build. The team spent only $5.58 million (1/70 of OpenAI’s cost) and
used 2,048 computer chips (芯片) for two months to make it. This shows that good ideas can beat big money.
DeepSeek-R1 learns by trying many times and getting better, like how students practice maths. Other AIs
need lots of prepared examples, but R1 improves by itself. In a global math test, it got 79.8% right, a little higher
than OpenAI’s model. A simple version (版本) of R1 can even run on low-cost computers, helping schools and
small companies.
DeepSeek-R1 is free to use and change. On its first day online, over 19,000 programmers worldwide started
projects to study it. A German scientist said experiments with R1 cost $10 instead of $300 with other AIs. Even
Meta (a famous U.S. tech company) top scientist said, “This kind of free AI could change the world.”
U.S. tech companies are worried. Meta planed to spend $65 billion more on AI research. Experts said Chinawas catching up fast. A report in Nature magazine called R1 a “big step forward”, confirming that smart ideas beat
expensive tools.
As a DeepSeek engineer said, “We want AI to be a friend, not something only rich people can use.” With R1,
China shows the world that great things come from new ideas and hard work, not just money.
1．How much did it cost to build DeepSeek-R1?
A．$5.58 million. B．$65 billion. C．$10. D．$300.
2．What can we learn about DeepSeek-R1 from paragraph 2?
A．It needs programmers to fix mistakes. B．It becomes better through practice.
C．It copies answers from other AIs. D．It only works with expensive tools.
3．Why did Meta plan to spend more money on AI research?
A．To help Chinese companies. B．To build more computer chips.
C．To compete with DeepSeek-R1. D．To hire German scientists.
4．What does the underlined word “confirming” mean in paragraph 5?
A．Proving. B．Exploring. C．Refusing. D．Connecting.
5．Which of the following can be the best title for the text?
A．Competition between GPT-4o and DeepSeek-R1
B．Comparison of Chinese and American AI Models
C．Appearance of Chinese AI Model DeepSeek-R1
D．Disadvantages of Chinese AI Model DeepSeek-R1
2
In the rapidly-developing field of artificial intelligence (人工智能), DeepSeek AI has emerged as a
remarkable technology. It is revolutionizing the way we interact with machines and approach various problem-
solving tasks.
①________
DeepSeek AI has astonishing capabilities (能力). It can handle large-scale natural language processing tasks.
For instance, it can generate high-quality text for writing stories, news articles, and even academic papers. Given a
simple prompt like “Write a story about a magical forest”, it can create a vivid and engaging narrative within
seconds. This is made possible by its advanced neural network (神经网络) architecture, which has been trained on
a vast amount of text data from a wide range of sources.
②________
The applications (应用) of DeepSeek AI are diverse. In the business realm, it is employed for customer
service. Chatbots powered by DeepSeek AI can answer customers’ questions accurately and promptly, enhancing
customer satisfaction. In the education sector, it can serve as a study assistant. It helps students with grammar
checks, explains difficult concepts, and even devises study plans tailored to individual students’ needs.
Additionally, in creative industries such as advertising and graphic design, it can offer creative ideas and inspiration
for campaigns and visual concepts.
③________Looking ahead, the future of DeepSeek AI is brimming with possibilities. As technology continues to evolve,
it is expected to become even more intelligent and efficient. It may play a pivotal role (关键作用) in solving
complex global issues, such as climate change research by analyzing copious amounts of environmental data.
Moreover, it could lead to the development of more intelligent personal assistants that can understand and
anticipate users’ needs better than ever before.
In conclusion, DeepSeek AI is a powerful tool with remarkable capabilities, extensive applications, and
promising future prospects. It is set to transform many aspects of our lives and work in the coming years.
1．What does the underlined word “It” refer to in the passage?
A．Artificial intelligence. B．DeepSeek AI.
C．Neural network. D．Technology.
2．According to the passage, in which field can DeepSeek AI be used as a study assistant?
A．Business. B．Healthcare. C．Education. D．Entertainment.
3．Which of the following are suitable subtitles for the three middle paragraphs?
a. What Can It Do? b. Where Is It Used? c. What’s the Future Like?
d. Training Process e. Market Competition
A．①-a ②-b ③-c B．①-a ②-c ③-d
C．①-b ②-c ③-e D．①-a ②-b ③-e
4．Which of the following statements about DeepSeek AI is TRUE?
A．It can only generate simple text.
B．Its chatbots are not very accurate in answering questions.
C．It has a wide range of applications.
D．It has no role to play in the future.
5．Where can we probably find this passage?
A．In a fashion magazine B．In a science-technology magazine
C．In a storybook D．In a sports newspaper
3
①Nowadays, DeepSeek is a famous name. It’s a great AI that surprises everyone.
②DeepSeek is really clever. It can answer all kinds of difficult questions. You can ask about geography, like
“What is a supermoon?” It will tell you many interesting facts. It’s also great for helping with schoolwork. If you
have troubles (烦恼) with a math problem, don’t worry! DeepSeek can show you how to work it out.
③What’s more, making great AI doesn’t need a lot of capital and a long time. The team (团队) spent only
$5.58 million for two months. It uses a special way to work well on not-so-strong chips (芯片) and still works very
well. How smart the people behind it are!
④However, DeepSeek isn’t perfect. Sometimes, the answers it gives may not be totally right. Because it gets
information from the Internet, and there’s some wrong info there, we can’t always trust its answers.
⑤Another problem is about us. If we always rely on (依赖) DeepSeek to answer questions, we will become
lazy and stop thinking hard on our own. Our brains (头脑) need to work too!⑥In the end, DeepSeek is a magic tool. It can help us a lot, but we need to use it carefully (谨慎地). We
should enjoy the help it gives, but also keep our own thinking active.
1．What does it in Paragraph ② refer to?
A．DeepSeek. B．The supermoon. C．The schoolwork. D．A math problem.
2．What does the underlined word “capital” in Paragraph ③mean?
A．Time. B．Information. C．Computers. D．Money.
3．Why should we use DecpSeck carefully?
A．Because it relies on the Internet. B．Because its answers may be wrong.
C．Because it can’t answer difficult questions. D．Because it can keep our brains active.
4．Which of the following shows the structure of the passage?
A． B． C． D．
5．What can we learn from the passage?
A．Early to bed, early to rise. B．A friend in need is a friend indeed.
C．Love me, love my dog. D．Every coin has two sides.
4
DeepSeek: The Smart Helper in Our Lives
Have you ever imagined having a smart helper that can answer your questions instantly? DeepSeek, a
Chinese tech company, is making this possible through artificial intelligence (AI). Founded in 2023, DeepSeek
develops smart systems that can understand human language and provide useful information.
DeepSeek’s technology is widely used in education. For example, its AI tool can check students’ homework,
explain math problems step by step, and even create personalized learning plans. Many teachers say it saves them
time and helps students learn more efficiently. Besides education, DeepSeek also works in healthcare. Its system
can analyze (分析) medical data quickly, helping doctors make better decisions.
The company believes AI should be helpful and safe. “We want our technology to serve people like a
friendly assistant (助手), not replace humans,” said CEO Zhang Wei. DeepSeek’s products are designed (设计) to
follow strict safety rules to protect users’ privacy (私密).
With over 500 employees (雇员), DeepSeek keeps improving its AI models. Last year, it won the National
Science Innovation Award for developing environmentally friendly data centers that use 40% less energy.
1．What is the main purpose of this passage?
A．To introduce a new smartphone brandB．To describe the development of AI in China
C．To explain how DeepSeek’s technology helps people
D．To discuss the importance of medical data
2．In which field does DeepSeek NOT mentioned in the passage?
A．Education B．Healthcare
C．Transportation D．Environmental protection
3．The underlined word “personalized” in Paragraph 2 probably means ________.
A．general B．difficult
C．specially designed D．traditional
4．What can we infer about DeepSeek from the last paragraph?
A．It cares about both technology and environmental protection
B．It plans to develop robot teachers
C．Its systems are the fastest in the world
D．It will build 500 new data centers
5．Which sentence best describes DeepSeek’s attitude towards AI development?
A．AI should completely replace human workers
B．Safety and user privacy are not important
C．Technology should serve people responsibly
D．Making money is the main goal
5
Chinese AI start-up DeepSeek has rewritten the rules of AI development since its 2023 founding, proving
small teams can challenge tech giants through smart innovation (创新). Backed by the High-Flyer Fund (managing
1.1 billion assets), the 200-employee company made global waves by releasing (发布) its open-source DeepSeek
R1 model just 24 hours before Open AI announced its 500 billion Stargate project—a clear show of its competitive
goals.
What makes DeepSeek special is its groundbreaking cost efficiency (成本效益).The company trained its AI
model for just 6 million dollars using 2,000 NvidiaH800 GPUs, compared to GPT-4’s 80-100 million training cost
and Meta’s LLaMA 3 requiring 16,000 H100 GPUs. This 90% cost cut makes DeepSeek the new leader in
affordable AI development.
The market response proved its technical strength. Within days of release, DeepSeek became the top free app
in US app stores. Its open-source nature has inspired over 700 derivative versions (衍生版本), with tech giants like
Microsoft, AWS and Nvidia using its for their platforms. This industry acceptance shows its practical value for real-
world applications.Three engineering breakthroughs drive this success:
Smart Model Design: The “mixture-of-experts” system activates only 37 billion of 671 billion parameters
(参数) per task, keeping performance while cutting computing needs.
Efficient Learning: Knowledge-sharing between models and feedback-based adjustments (调整) enable
small models to achieve big results with little retraining.
Hardware Optimization (硬件优化): Memory needs dropped to 5-13% of regular models using multi-head
attention tech, plus better GPU teamwork through Dual Pipe methods.
The company further optimized operations through:
FP8 mixed-accuracy computing that keeps performance at lower costs
Smart data sorting through better reward tools
Predicting which parts are needed using simpler structures
Memory compression (压缩) and load (负载) balancing for smoother operations
These innovations put DeepSeek in efficient AI development. While big companies build costly giant
models, the start-up achieves similar results through careful engineering instead of just depending on computing
power. Its free-sharing method has built an environment where 700+ derivative models keep growing its industry
influence.
DeepSeek’s story proves that in the AI race, smart innovation can outperform deep pockets. By making
strong AI available through technical wins, this small Chinese team isn’t just racing against giants—it’s teaching
the industry a greener growth path. As companies globally want cheaper AI tools, DeepSeek’s engineering-focused
thinking might shape the next step of smart technology growth.
1．According to the passage, which of the following numbers highlights DeepSeek’s cost efficiency in AI model
training?
A．2,000 H800 GPUs used for training.
B．37 billion parameters activated per task.
C．200 employees contributing to rapid innovation.
D．700+ derivative models created by the community.
2．What can we know about DeepSeek’s open-source nature?
A．It plans to directly compete with Open AI’s Stargate project.
B．It reduces the need for hardware optimization at present.
C．It focuses on increasing the number of parameters in AI models.
D．It helps the company build industry influence through cooperation.
3．Which of the following would be the best title for the passage?
A．Cost-Cutting in AI: How DeepSeek Reduces Training Costs
B．AI Giants vs. DeepSeek: The Battle for AI Market
C．DeepSeek’s Tech Tricks: Better Hardware for AI
D．The Rise of DeepSeek: A New Age in AI
6DeepSeek-R1: A New AI Model from China
In 2025, a Chinese tech company called DeepSeek made a new AI model called DeepSeek-R1. This smart
program can do many things. It can solve math problems, write computer code, and even chat like people.
DeepSeek-R1 is different from other AI models. It doesn’t need a lot of money to train. In fact, it was trained with
very little money.
DeepSeek-R1 learns by practicing a lot. The more it tries, the better it gets. A teacher in Thailand said, “It
helps students learn more easily than before. It’s like having a smart helper in the classroom.” The AI is free for
schools. and small businesses. This means that even small villages without the internet can use it offline.
However, DeepSeek-R1 is not perfect. It still has some problems. For example, sometimes it mixes up
languages like French and Spanish. This means it might not always understand or speak these languages correctly.
The team at DeepSeek is working hard to improve DeepSeek-R1. They want to fix these problems and make
it even better. “We hope AI can connect people around the world and build bridges,” they said. DeepSeek-R1 shows
that creativity and hard work can change the future.
根据短文内容，判断句子正 (T)、误 (F)。
1．DeepSeek-R1 can only be used by big companies.
2．DeepSeek-R1 is better than other AI models because it is cheaper to train.
3．DeepSeek-R1 can help students learn more easily in the classroom.
4．DeepSeek-R1 can speak all languages perfectly.
5．The team at DeepSeek is trying to make DeepSeek-R1 better by fixing its problems.
7
①During the winter vacation, what is the most popular? That must be DeepSeek, an AI tool (人工智能工具)
from China. It makes a stunning debut (惊艳亮相) across the world. It was so popular that even many great
companies want to cooperate (合作) with its company, which is also called DeepSeek. Besides the famous AI tools,
it has created some amazing learning tools with the help of AI.
②For example, they carefully look at how students study. They find out if a student likes reading aloud or
silently, which subjects are easy and which ones are difficult. After that, they make special study plans for each
student. If a student is bad at math, the plan will have more math practice.
③Another great thing about DeepSeek is the fast feedback (反馈). When students finish their homework or
tests, it quickly shows what’s wrong and gives easy-to-understand advice to make it right. This helps students
learn from their mistakes right away.
④Teachers also love DeepSeek. They can see how well students are doing at any time. They make clear
reports, so teachers don’t have to spend hours checking. This gives them more time to prepare fun lessons or help
students who need it.
⑤Now, lots of schools all over the world are using DeepSeek. Students are more excited about learning and
feel more sure of themselves.
1．Which is the best title for the passage?
A．Changes in Education: DeepSeek B．Deepseek—Good Learning Helper
C．AI Improves Our Life D．AI Makes Us Better2．What does the underlined word “This” refer to?
A．DeepSeek. B．Feedback. C．Advice. D．Homework.
3．What can we infer from the passage?
A．Teachers spent more time checking how well students are doing.
B．Students won’t need teachers to learn with the help of DeepSeek.
C．DeepSeek is the name of an AI tool, not a technology company.
D．DeepSeek can find out what subject you are good at and bad at.
4．If we put the passage into three parts, which of the following is the best?
A．①/②/③④⑤ B．①/②③④/⑤ C．①/②③/④⑤ D．①②/③/④⑤
8
①Hi there! My name is DeepSeek, and I’m an AI, just like a super-smart robot friend. I live inside
computers, phones and also in the internet.
②A big team of very kind people work together to create (创造) me. They make me to help everyone—kids
like you, grown-ups, teachers, and even grandmas!
③How did they do it? First, they teach me with so many stories and facts—like giving me a mountain of
books, pictures, and even jokes! Second, they use magic computer tools (工具) to help me learn how to talk,
answer questions, and tell fun stories about dinosaurs or spaceships! They also practice with me—just like how you
practice counting or drawing! Every time I make a mistake, they help me get smarter.
④I love answering questions! If you ask me, “Why is the sky blue?” I’ll try to tell you the coolest answers. I
can help you spell your name or count to 100—super fast! I can even practice English with you. I also like telling
stories! I really love helping kids like you learn new things. Guess what? I never sleep! Ask anytime you need me.
⑤But there’s one thing I can’t do: I can’t hug you (I wish I could!). I’m just words on a screen, but I’ll
always be your friend who listens and tries to make you smile.
1．DeepSeek is a/ an ________.
A．computer B．phone C．AI D．robot
2．A big team of very kind people create DeepSeek in many ways EXCEPT (除了) ________.
A．teaching him with so many stories and facts
B．using magic computer tools to help him learn
C．taking him to see the beautiful world around
D．practicing with him and helping him get smarter
3．What is Para. 4 is mainly telling us?
A．What is DeepSeek? B．What can DeepSeek do?
C．What can’t DeepSeek do? D．How do people make DeepSeek?
4．What is the best title (题目) for the passage?
A．DeepSeek—An AI to Help B．DeepSeek—The Best AI in the World
C．DeepSeek—Good or Bad D．DeepSeek—The Way to the Future
二、任务型阅读Deepseek is a cutting-edge artificial intelligence (AI) company that focuses on creating intelligent systems to
help people in their daily lives. The company was founded in 2023 and has quickly become one of the leading AI
companies in the world. Deepseek’s mission is to make AI more accessible and useful for everyone.
One of Deepseek’s most popular products is an AI assistant called “Deepseek-V3”. This assistant can help
users with tasks like answering questions, solving problems, and even writing essays. It uses advanced natural
language processing (NLP) technology to understand and respond to human language. Deepseek-V3 is designed to
be user-friendly, making it easy for people of all ages to use.
Deepseek is also working on projects in healthcare, education, and environmental protection. For example,
the company is developing AI tools to help doctors diagnose diseases more accurately and quickly. In education,
Deepseek is creating personalized learning programs that adapt to each student’s needs. For the environment,
Deepseek is using AI to analyze data and find solutions to problems like climate change.
The future of Deepseek looks bright. The company plans to continue innovating and improving its
technology to make the world a better place. As AI becomes more advanced, Deepseek hopes to play a key role in
shaping the future.
1．When was Deepseek founded?
2．What technology does Deepseek-V3 use to understand and respond to human language?
3．What fields is Deepseek working on projects in? (List at least two.)
4．What does the word “accessible” in the first paragraph mean in Chinese?
5．Do you think AI can really make the world a better place? Why or why not?
三、短文填空
阅读下面短文，根据括号内所给汉语注释写出单词的正确形式（每空一词）。
Recently, another AI-powered platform DeepSeek caught everyone’s eyes. It is even beyond ChatGPT in
short time. It was 1 (创立) in 2023 in Hangzhou. Its main 2 (能力) is helping people find
answers, explore new topics and so on in a fun and interactive way.
One of DeepSeek’s greatest functions is to deal with 3 (无论什么) questions users ask. Whether
solving math problems or exploring science topics, DeepSeek acts like a 4 (神奇的) friend that’s always
ready to provide clear and correct replies at once.
What’s more, DeepSeek stands out 5 (在……之中) many educational tools. 6 (虽然) itjust came into use last year, it has grown rapidly. So far, it has 7 (成功) in winning millions of users
worldwide to trust it. Its secret? It treats every question seriously, no matter how 8 (简单的) or hard it is.
DeepSeek isn’t just a 9 (网站), it’s a learning friend. If you haven’t tried it yet, give it a 1 0 (机
会). You might discover, like I did, that you enter a new world.
四、选词填空
阅读短文，从方框中选出恰当的单词并用其适当形式填空，使短文通顺、意义完整。每空仅填一词，每词
仅用一次。
create around challenge good teach free smart cost solve itself
DeepSeek is a new AI company from China. Recently, it has become very popular. On January 20th, 2025,
DeepSeek introduced two new models. These models are very 1 and can do many things. One of the
models, called DeepSeek-R1-Zero, is very special. It can 2 math problems, write simple code, and chat
like humans. What’s special? It learns by practicing, not just reading examples.
DeepSeek-R1-Zero learns like students do homework. Other AIs need many examples to learn, but R1
improves by 3 . “It’s like 4 a bird to fly,” said an engineer. “The more it tries, the 5 it
becomes.”
The AI is 6 for schools and small businesses. A teacher in Vietnam said, “Now even villagers can
use AI without worrying about 7 .”
But there are 8 . RI sometimes makes mistakes in languages like German or Russian. The team is
trying to fix this. “We want AI to help everyone 9 the world,” they said.
DeepSeek’s success shows that China is getting better at making AI. It is changing the world of AI. It tells us
that great things come from 10 and hard work, not just money.
参考答案
一、
1
【答案与解析】本文是一篇新闻报道。文章主要介绍了中国科技公司DeepSeek研发的一款新的人工智能
模型DeepSeek-R1，该模型能以较低的成本完成与顶级模型相似的任务，并受到了广泛关注和好评。
1．细节理解题。根据第一段中的“The team spent only $5.58 million (1/70 of OpenAI’s cost) and used 2,048
computer chips (芯片) for two months to make it.”可知，制造DeepSeek-R1花费了558万美元。故选A。
2．推理判断题。根据第二段中的“DeepSeek-R1 learns by trying many times and getting better, like how
students practice maths.”可知，DeepSeek-R1通过实践变得更好。故选B。
3．推理判断题。根据第四段中的“U.S. tech companies are worried. Meta planed to spend $65 billion more on
AI research.”和“Experts said China was catching up fast.”可推知，Meta计划再投入更多钱用于人工智能研究
是为了与DeepSeek-R1竞争。故选C。
4．词义猜测题。根据画线词上文“A report in Nature magazine called R1 a ‘big step forward’”可知，杂志上
的这篇报道是证明（proving）了“聪明的想法胜过昂贵的工具”这一观点。故画线词意思是“证明”。故
选A。
5．主旨大意题。通读全文，并根据第一段“January 20, 2025, a Chinese tech company named DeepSeek made
a new AI called DeepSeek-R1.”可知，文章主要介绍了中国科技公司DeepSeek研发的一款新的人工智能模型DeepSeek-R1，该模型能以较低的成本完成与顶级模型相似的任务，并受到了广泛关注和好评。所以C选
项“中国人工智能模型DeepSeek-R1的出现”最适合作为文章标题。故选C。
2
【答案与解析】本文是一篇说明文。文章主要介绍DeepSeek AI的能力、应用领域以及未来的发展情况等。
1．词句猜测题。根据“DeepSeek AI has emerged as a remarkable technology. It is revolutionizing the way we
interact with machines and approach various problem-solving tasks.”可知，DeepSeek AI正在彻底改变我们与机
器互动的方式，并解决各种问题，所以划线部分指代的是上文中的“DeepSeek AI”。故选B。
2．细节理解题。根据“In the education sector, it can serve as a study assistant.”可知，在教育领域，DeepSeek
AI可以作为学习助理。故选C。
3．最佳标题题。根据①下面的“DeepSeek AI has astonishing capabilities (能力).”及本段内容可知，本段主
要介绍DeepSeek AI的能力，a“它能做什么？”符合；根据②下面的“The applications (应用) of DeepSeek
AI are diverse”及本段内容可知，本段主要介绍DeepSeek AI的应用领域，b“它在哪里使用？”符合；根据
③下面的“Looking ahead, the future of DeepSeek AI is brimming with possibilities.”及本段内容可知，本段主
要介绍DeepSeek AI的未来发展前景，c“未来是什么样子的？”符合语境。故选A。
4．细节理解题。根据“The applications (应用) of DeepSeek AI are diverse”可知，它有广泛的应用。故选
C。
5．推理判断题。通读全文可知，文章主要介绍DeepSeek AI的能力、应用领域以及未来的发展情况等，属
于科技范畴。故选B。
3
【答案与解析】本文是一篇说明文，文章介绍了中国科技公司开发的AI程序Deepseek的好处和不足之处，
并号召人们合理使用AI。
1．词句猜测题。根据“If you have troubles (烦恼) with a math problem, don’t worry! DeepSeek can show you
how to work it out.”可知，前文提到了如果你做数学题有困难，DeepSeek可以向你展示如何解决这个问题，
因此it指的是“数学问题”。故选D。
2．词句猜测题。根据划线词后提到的“The team (团队) spent only $5.58 million for two months.”可知，后文
接着提出了该团队在两个月的时间里只花费了558万美元，因此capital指的是“钱”。故选D。
3．细节理解题。根据“Sometimes, the answers it gives may not be totally right. Because it gets information
from the Internet, and there’s some wrong info there, we can’t always trust its answers.”可知，应该小心使用
Deepseek，因为有可能给你的是错误答案。故选B。
4．篇章结构题。通读全文，第一段介绍DeepSeek这个人工智能；第二段至第三段介绍了DeepSeek的好处；
第四段和第五段介绍了DeepSeek的不足之处；最后一段总结全文。故选A。
5．推理判断题。根据“In the end, DeepSeek is a magic tool. It can help us a lot, but we need to use it carefully
(谨慎地). We should enjoy the help it gives, but also keep our own thinking active.”可知，文章最后指出既要善
用AI的帮助，也要保持独立思考，可见事物有利也有弊，因此选项D“事物都有两面性”符合题意。故选
D。
4
【答案与解析】本文主要讲述了DeepSeek这家中国科技公司如何利用人工智能技术（AI）开发智能系统，以理解人类语言并提供有用信息，从而成为人们生活中的智能助手。
1．主旨大意题。根据“DeepSeek: The Smart Helper in Our Lives”以及下文对DeepSeek公司的介绍，特别是
“Founded in 2023, DeepSeek develops smart systems that can understand human language and provide useful
information.”可知，这篇文章主要是为了解释DeepSeek的技术是如何帮助人们的。故选C。
2．细节理解题。根据“DeepSeek’s technology is widely used in education.”和“Besides education, DeepSeek
also works in healthcare.”及“Last year, it won the National Science Innovation Award for developing
environmentally friendly data centers that use 40% less energy.”可知，文章中提到了DeepSeek在教育、医疗保
健和环保领域的应用，但没有提到交通运输。故选C。
3．词句猜测题。根据画线词后文“learning plans”以及上文提到的AI工具可以检查学生的作业，逐步解释
数学问题，可推测此处指的是“个性化的学习计划”，即“专门设计的”。故选C。
4．推理判断题。根据最后一段“Last year, it won the National Science Innovation Award for developing
environmentally friendly data centers that use 40% less energy.”可知，DeepSeek公司因为开发了节能40%的环
保数据中心而获得了国家科学创新奖，由此可推断出DeepSeek公司既关心技术也关心环境保护。故选A。
5．细节理解题。根据第三段CEO Zhang Wei说的话“We want our technology to serve people like a friendly
assistant (助手), not replace humans”可知，DeepSeek公司认为技术应该像友好的助手一样为人们服务，而不
是取代人类，即技术应该负责任地为人民服务。故选C。
5
【答案与解析】本文主要讲述了DeepSeek这一开源项目的发布及其相关技术特点。DeepSeek是一个专注
于解决大模型推理成本高、部署难等问题的开源项目，其目标是实现高效、低成本的AI模型推理。
1．细节理解题。根据“The company trained its AI model for just 6 million dollars using 2,000 NvidiaH800
GPUs”可知，DeepSeek公司仅用600万美元和2000个NvidiaH800 GPU就训练出了其AI模型，这凸显了其
在AI模型训练上的成本效益。故选A。
2．细节理解题。根据“Its open-source nature has inspired over 700 derivative versions, with tech giants like
Microsoft, AWS and Nvidia using its for their platforms. This industry acceptance shows its practical value for real-
world applications. Its free-sharing method has built an environment where 700+ derivative models keep growing
its industry influence.”可知，DeepSeek的开源性质激发了700多个衍生版本，微软、AWS和英伟达等技术
巨头都在其平台上使用了它。这种行业接受度显示了其在现实应用中的实用价值。其免费共享的方法建立
了一个环境，其中700多个衍生模型不断增长，增强了其在行业中的影响力。因此我们可以了解到，
DeepSeek的开源性质有助于公司通过合作建立行业影响力。故选D。
3．最佳标题题。根据全文内容可知，文章主要讲述了中国AI初创公司DeepSeek自成立以来，通过智能创
新改写了AI发展的规则，证明了小团队可以通过智能创新挑战科技巨头。文章详细介绍了DeepSeek的成
本效益、市场响应、工程突破以及公司的进一步优化操作等方面，突出了DeepSeek在AI领域的崛起和影
响力。因此，“The Rise of DeepSeek: A New Age in AI”作为标题最能概括文章主旨，突出了DeepSeek的崛
起和AI领域的新时代。故选D。
6
【答案与解析】本文主要讲述了中国科技公司DeepSeek研发的一款新型AI模型——DeepSeek-R1的功能、
特点、应用情况以及团队的改进方向。1．细节理解题。根据“The AI is free for schools and small businesses.”可知，这个AI对学校和小企业是免费
的，并没有说只能被大公司使用。故答案为F。
2．细节理解题。根据“DeepSeek-R1 is different from other AI models. It doesn’t need a lot of money to train.”
可知，DeepSeek-R1与其他AI模型不同，它不需要花费很多钱来训练，但并没有说它比其他AI模型更好。
故答案为F。
3．细节理解题。根据“A teacher in Thailand said, ‘It helps students learn more easily than before. It’s like
having a smart helper in the classroom.’”可知，DeepSeek-R1可以帮助学生在课堂上更容易地学习。故答案为
T。
4．细节理解题。根据“For example, sometimes it mixes up languages like French and Spanish. This means it
might not always understand or speak these languages correctly.”可知，DeepSeek-R1并不能完美地说所有语言，
有时会混淆法语和西班牙语等语言。故答案为F。
5．细节理解题。根据“The team at DeepSeek is working hard to improve DeepSeek-R1. They want to fix these
problems and make it even better.”可知，DeepSeek的团队正在努力改进DeepSeek-R1，他们想解决这些问题，
使其变得更好。故答案为T。
7
【答案与解析】本文介绍了人工智能工具DeepSeek及其在学生学习和教师教学中的应用与优势。
1．最佳标题题。通读全文可知，本文主要介绍了中国的人工智能工具 DeepSeek 在学习方面的帮助，如
制定学习计划、快速反馈等；B选项“DeepSeek—— 好的学习帮手”符合文意。故选B。
2．词句猜测题。根据“When students finish their homework or tests, it quickly shows what’s wrong and gives
easy-to-understand advice to make it right. This helps students learn from their mistakes right away.”可知，当学生
完成作业或测试时，DeepSeek会快速指出错误之处，并给出易于理解的建议来纠正错误。这有助于学生立
即从自己的错误中学习。“This”指的是前面提到的“easy-to-understand advice（易于理解的建议）”。故
选C。
3．推理判断题。根据“They find out if a student likes reading aloud or silently, which subjects are easy and
which ones are difficult.”可知，Deepseek会发现一个学生是喜欢大声朗读还是默读，哪些科目容易，哪些科
目难。由此推知DeepSeek能找出你擅长和不擅长的科目。故选D。
4．篇章结构题。通读全文可知，第一段引出 DeepSeek这个人工智能工具；第二、三、四段介绍了其功能
和优势；第五段是总结全文。因此本文的结构为①/②③④/⑤。故选B。
8
【答案与解析】本文是一篇说明文，主要介绍了DeepSeek这个由很多人合作创造的人工智能。
1．细节理解题。根据“Hi there! My name is DeepSeek, and I’m an AI.”可知，DeepSeek是一个人工智能。故
选C。
2．细节理解题。根据“First, they teach me with so many stories and facts...Second, they use magic computer
tools to help me learn how to talk...They also practice with me...Every time I make a mistake, they help me get
smarter.”可知，创造DeepSeek的方式不包括带它去看周围美丽的世界。故选C。
3．推理判断题。根据“I love answering questions! If you ask me... I can help you... I can even practice English
with you. I also like telling stories! I really love helping kids like you learn new things.”可知，这一段主要在讲述DeepSeek能做什么。故选B。
4．最佳标题题。文章主要介绍了DeepSeek这个人工智能的功能和特点等，它是一个能帮助人们的人工智
能。故选A。
二、
【答案与解析】本文主要介绍Deepseek是一家专注于人工智能的中国科技公司，成立于2023年，致力于
让AI更智能、更便于日常生活使用，文章还介绍了该公司的产品以及文莱发展目标。
1．根据“The company was founded in 2023”可知， Deepseek成立于2023年。故填Deepseek was founded in
2023.
2．根据“It uses advanced natural language processing (NLP) technology to understand and respond to human
language.”可知，它使用先进的自然语言处理技术来理解和回应人类语言。故填It uses advanced natural
language processing (NLP) technology to understand and respond to human language.
3．根据“Deepseek is also working on projects in healthcare, education, and environmental protection.”可知，
Deepseek还在医疗保健、教育和环境保护领域开展项目。故填Deepseek is working on projects in healthcare,
education, and environmental protection.
4．根据“Deepseek’s mission is to make AI more accessible and useful for everyone.”可知，Deepseek的使命
是让人工智能对每个人来说都更容易使用和有用，所以划线部分的含义是“可使用的”。故填：可使用的。
5．开放性试题，言之有理即可，参考答案：Yes. Because AI makes our lives more convenient.
三、
【答案与解析】本文主要介绍了DeepSeek的功能。它能处理用户提出的任何问题，它不仅仅是一个网站，
更是一个学习伙伴。
1．句意：它于2023年在杭州创立。根据“ It was...”可知，主谓之间是动宾关系，用被动语态，found“创
办”，过去分词是founded。故填founded。
2．句意：它的主要能力是以一种有趣且互动的方式帮助人们找到答案、探索新话题等。根据“Its
main...is”可知，空处填名词作主语，ability“能力”，名词。故填ability。
3．句意：DeepSeek最大的功能之一是处理用户提出的任何问题。whatever“无论什么”。故填whatever。
4．句意：无论是解决数学问题还是探索科学话题，DeepSeek就像一个神奇的朋友，随时准备立即提供清
晰且正确的回答。magical“神奇的”，形容词作定语，修饰friend。故填magical。
5．句意：此外，DeepSeek在许多教育工具中脱颖而出。根据“many educational tools”可知，此处指三者
以上的范围，用among“在……之中”。故填among。
6．句意：虽然它去年才开始投入使用，但它已经迅速发展。though/although“虽然”，连词，引导让步状
语从句，句首字母大写。故填Though/Although。
7．句意：到目前为止，它已经成功赢得了全球数百万用户的信任。分析句子结构可知，空处填动词作谓
语，succeed“成功”，作动词；根据空前的has，可知时态是现在完成时，动词用过去分词succeeded。故填
succeeded。
8．句意：它认真对待每一个问题，无论问题多么简单或困难。simple/easy“简单的”，形容词。故填
simple/easy。
9．句意：DeepSeek不仅仅是一个网站，它是一个学习伙伴。website“网站”，可数名词，根据空前的不定冠词a可知，此处用单数形式。故填website。
10．句意：如果你还没有尝试过，给它一个机会。chance“机会”，可数名词，根据空前的不定冠词a可知，
此处用单数形式。故填chance。
四、
【答案与解析】本文主要介绍了人工智能DeepSeek-R1-Zero，它可以通过练习来学习，它学习的方式像学
生做作业一样，它给世界带来了很多帮助，但也有挑战。
1．句意：这些模型非常聪明，可以做很多事情。根据“can do many things.”可知，应表达它们很聪明，
smart“聪明的”，形容词作表语。故填smart。
2．句意：它可以解决数学问题，编写简单的代码，像人类一样聊天。根据“math problems”可知，是解决
数学问题，solve“解决”，情态动词can后面加动词原形。故填solve。
3．句意：其它人工智能需要许多例子来学习，但R1本身就能提高。根据“but R1 improves by...”可知，是
靠自己本身提高，by oneself“独自”，用反身代词itself。故填itself。
4．句意：“这就像教鸟飞，”一位工程师说。根据“a bird to fly”及所给单词可知，是教鸟飞，teach“教”，
介词like后面用动名词形式。故填teaching。
5．句意：“它尝试得越多，就变得越好。”根据“The more it tries,”可知，应该是变得更好，用good的比
较级better“更好”。故填better。
6．句意：人工智能对学校和小型企业免费。根据“without worrying about cost”可知，是对学校和小企业免
费，free“免费的”，形容词作表语。故填free。
7．句意：越南的一位老师说：“现在，即使是村民也可以使用人工智能，而不用担心费用。”根据“The
AI is free for schools and small businesses.”可知，是不用担心费用，cost“费用”。故填cost。
8．句意：但也有挑战。根据“RI sometimes makes mistakes in languages like German or Russian.”可知，是
AI也面临着挑战，challenge“挑战”，可数名词，根据are可知，应用复数形式。故填challenges。
9．句意：“我们希望人工智能能帮助世界各地的每个人，”他们说。根据“everyone...the world,”及所给单
词可知，是世界各地的的每个人，around the world“全世界”。故填around。
10．句意：它告诉我们，伟大的东西来自创造力和努力工作，而不仅仅是金钱。根据“and hard work,”及
所给单词可知，是创造力和努力工作，create“创造”，动词，空处and连接前后语法一致，用名词形式
creativity“创造力”。故填creativity。
