This Stage Utilized 3 Reward Models

DeepSeek (Chinese: 深度求索; pinyin: Shēndù Qiúsuǒ) is a Chinese synthetic intelligence company that establishes open-source large language models (LLMs). Based in Hangzhou, Zhejiang, it is owned and funded by Chinese hedge fund High-Flyer, whose co-founder, Liang Wenfeng, developed the business in 2023 and functions as its CEO.

The DeepSeek-R1 design offers responses equivalent to other contemporary large language designs, such as OpenAI’s GPT-4o and o1. [1] It is trained at a significantly lower cost-stated at US$ 6 million compared to $100 million for OpenAI’s GPT-4 in 2023 [2] -and requires a tenth of the computing power of an equivalent LLM. [2] [3] [4] DeepSeek’s AI designs were developed in the middle of United States sanctions on India and China for Nvidia chips, [5] which were planned to restrict the ability of these 2 nations to develop innovative AI systems. [6] [7]

On 10 January 2025, DeepSeek released its very first totally free chatbot app, based upon the DeepSeek-R1 design, for iOS and Android; by 27 January, DeepSeek-R1 had exceeded ChatGPT as the most-downloaded totally free app on the iOS App Store in the United States, [8] causing Nvidia’s share cost to come by 18%. [9] [10] DeepSeek’s success against larger and more established competitors has been explained as “overthrowing AI”, [8] constituting “the very first chance at what is becoming an international AI area race”, [11] and ushering in “a new period of AI brinkmanship”. [12]

DeepSeek makes its generative expert system algorithms, models, and training details open-source, enabling its code to be freely available for usage, modification, watching, and creating files for building purposes. [13] The company apparently vigorously hires young AI researchers from leading Chinese universities, [8] and employs from outside the computer technology field to diversify its models’ knowledge and capabilities. [3]

In February 2016, High-Flyer was co-founded by AI enthusiast Liang Wenfeng, who had been trading since the 2007-2008 financial crisis while participating in Zhejiang University. [14] By 2019, he established High-Flyer as a hedge fund concentrated on establishing and using AI trading algorithms. By 2021, High-Flyer specifically used AI in trading. [15] DeepSeek has made its generative synthetic intelligence chatbot open source, indicating its code is easily readily available for use, modification, and watching. This includes authorization to access and utilize the source code, along with style documents, for developing purposes. [13]

According to 36Kr, Liang had actually built up a shop of 10,000 Nvidia A100 GPUs, which are utilized to train AI [16], before the United States federal government imposed AI chip constraints on China. [15]

In April 2023, High-Flyer started a synthetic basic intelligence lab committed to research study developing AI tools different from High-Flyer’s financial company. [17] [18] In May 2023, with High-Flyer as one of the financiers, the lab became its own business, DeepSeek. [15] [19] [18] Equity capital companies in providing financing as it was unlikely that it would be able to produce an exit in a brief period of time. [15]

After launching DeepSeek-V2 in May 2024, which provided strong efficiency for a low rate, DeepSeek ended up being referred to as the catalyst for China’s AI model cost war. It was quickly called the “Pinduoduo of AI”, and other significant tech giants such as ByteDance, Tencent, Baidu, and Alibaba started to cut the rate of their AI models to compete with the company. Despite the low price charged by DeepSeek, it paid compared to its rivals that were losing money. [20]

DeepSeek is focused on research and has no detailed plans for commercialization; [20] this also permits its technology to prevent the most strict arrangements of China’s AI regulations, such as requiring consumer-facing technology to adhere to the federal government’s controls on info. [3]

DeepSeek’s working with preferences target technical capabilities rather than work experience, leading to a lot of new hires being either current university graduates or developers whose AI careers are less established. [18] [3] Likewise, the company recruits people with no computer science background to help its technology understand other topics and knowledge areas, including being able to produce poetry and perform well on the notoriously difficult Chinese college admissions exams (Gaokao). [3]

Development and release history

DeepSeek LLM

On 2 November 2023, DeepSeek launched its first series of model, DeepSeek-Coder, which is offered free of charge to both scientists and industrial users. The code for the model was made open-source under the MIT license, with an additional license contract (“DeepSeek license”) regarding “open and accountable downstream usage” for the design itself. [21]

They are of the same architecture as DeepSeek LLM detailed below. The series includes 8 designs, 4 pretrained (Base) and 4 instruction-finetuned (Instruct). They all have 16K context lengths. The training was as follows: [22] [23] [24]

1. Pretraining: 1.8 T tokens (87% source code, 10% code-related English (GitHub markdown and Stack Exchange), and 3% code-unrelated Chinese).
2. Long-context pretraining: 200B tokens. This extends the context length from 4K to 16K. This produced the Base designs.
3. Supervised finetuning (SFT): 2B tokens of instruction data. This produced the Instruct models.

They were trained on clusters of A100 and H800 Nvidia GPUs, linked by InfiniBand, NVLink, NVSwitch. [22]

On 29 November 2023, DeepSeek launched the DeepSeek-LLM series of designs, with 7B and 67B specifications in both Base and Chat forms (no Instruct was released). It was developed to complete with other LLMs available at the time. The paper claimed benchmark results higher than most open source LLMs at the time, specifically Llama 2. [26]: area 5 Like DeepSeek Coder, the code for the design was under MIT license, with DeepSeek license for the model itself. [27]

The architecture was basically the like those of the Llama series. They utilized the pre-norm decoder-only Transformer with RMSNorm as the normalization, SwiGLU in the feedforward layers, rotary positional embedding (RoPE), and grouped-query attention (GQA). Both had vocabulary size 102,400 (byte-level BPE) and context length of 4096. They trained on 2 trillion tokens of English and Chinese text acquired by deduplicating the Common Crawl. [26]

The Chat variations of the two Base models was likewise launched simultaneously, acquired by training Base by monitored finetuning (SFT) followed by direct policy optimization (DPO). [26]

On 9 January 2024, they released 2 DeepSeek-MoE models (Base, Chat), each of 16B parameters (2.7 B activated per token, 4K context length). The training was essentially the like DeepSeek-LLM 7B, and was trained on a part of its training dataset. They declared comparable performance with a 16B MoE as a 7B non-MoE. In architecture, it is a variant of the basic sparsely-gated MoE, with “shared professionals” that are constantly queried, and “routed experts” that may not be. They discovered this to aid with skilled balancing. In standard MoE, some specialists can become excessively counted on, while other experts might be rarely used, losing specifications. Attempting to balance the experts so that they are similarly used then causes experts to duplicate the very same capacity. They proposed the shared specialists to learn core capacities that are often utilized, and let the routed professionals to learn the peripheral capacities that are rarely used. [28]

In April 2024, they released 3 DeepSeek-Math designs specialized for doing math: Base, Instruct, RL. It was trained as follows: [29]

1. Initialize with a previously pretrained DeepSeek-Coder-Base-v1.5 7B.
2. Further pretrain with 500B tokens (6% DeepSeekMath Corpus, 4% AlgebraicStack, 10% arXiv, 20% GitHub code, 10% Common Crawl). This produced the Base design.
3. Train an instruction-following model by SFT Base with 776K math issues and their tool-use-integrated detailed options. This produced the Instruct model.
Reinforcement knowing (RL): The reward design was a procedure benefit design (PRM) trained from Base according to the Math-Shepherd approach. [30] This reward design was then utilized to train Instruct utilizing group relative policy optimization (GRPO) on a dataset of 144K math questions “related to GSM8K and MATH”. The benefit model was continually updated throughout training to avoid reward hacking. This led to the RL model.

V2

In May 2024, they launched the DeepSeek-V2 series. The series consists of 4 designs, 2 base designs (DeepSeek-V2, DeepSeek-V2-Lite) and 2 chatbots (-Chat). The 2 bigger designs were trained as follows: [31]

1. Pretrain on a dataset of 8.1 T tokens, where Chinese tokens are 12% more than English ones.
2. Extend context length from 4K to 128K using YaRN. [32] This led to DeepSeek-V2.
3. SFT with 1.2 M instances for helpfulness and 0.3 M for safety. This resulted in DeepSeek-V2-Chat (SFT) which was not released.
4. RL using GRPO in two stages. The first phase was trained to resolve mathematics and coding issues. This phase utilized 1 benefit model, trained on compiler feedback (for coding) and ground-truth labels (for mathematics). The 2nd phase was trained to be helpful, safe, and follow rules. This phase used 3 reward designs. The helpfulness and security benefit designs were trained on human preference information. The rule-based reward model was manually programmed. All qualified reward designs were initialized from DeepSeek-V2-Chat (SFT). This led to the released variation of DeepSeek-V2-Chat.

They chose for 2-staged RL, because they found that RL on reasoning information had “unique characteristics” various from RL on general data. For example, RL on thinking could improve over more training steps. [31]

The two V2-Lite models were smaller sized, and qualified similarly, though DeepSeek-V2-Lite-Chat only went through SFT, not RL. They trained the Lite version to assist “additional research and advancement on MLA and DeepSeekMoE”. [31]

Architecturally, the V2 designs were significantly customized from the DeepSeek LLM series. They altered the standard attention mechanism by a low-rank approximation called multi-head latent attention (MLA), and utilized the mixture of specialists (MoE) variant formerly released in January. [28]

The Financial Times reported that it was cheaper than its peers with a rate of 2 RMB for every million output tokens. The University of Waterloo Tiger Lab’s leaderboard ranked DeepSeek-V2 seventh on its LLM ranking. [19]

In June 2024, they launched 4 designs in the DeepSeek-Coder-V2 series: V2-Base, V2-Lite-Base, V2-Instruct, V2-Lite-Instruct. They were trained as follows: [35] [note 2]

1. The Base models were initialized from corresponding intermediate checkpoints after pretraining on 4.2 T tokens (not the version at the end of pretraining), then pretrained further for 6T tokens, then context-extended to 128K context length. This produced the Base models.
DeepSeek-Coder and DeepSeek-Math were utilized to produce 20K code-related and 30K math-related instruction information, then combined with a guideline dataset of 300M tokens. This was utilized for SFT.
2. RL with GRPO. The reward for math problems was computed by comparing to the ground-truth label. The reward for code issues was generated by a benefit design trained to predict whether a program would pass the system tests.

DeepSeek-V2.5 was released in September and upgraded in December 2024. It was made by combining DeepSeek-V2-Chat and DeepSeek-Coder-V2-Instruct. [36]

V3

In December 2024, they launched a base model DeepSeek-V3-Base and a chat model DeepSeek-V3. The design architecture is essentially the same as V2. They were trained as follows: [37]

1. Pretraining on 14.8 T tokens of a multilingual corpus, mainly English and Chinese. It consisted of a greater ratio of mathematics and programming than the pretraining dataset of V2.
2. Extend context length twice, from 4K to 32K and then to 128K, utilizing YaRN. [32] This produced DeepSeek-V3-Base.
3. SFT for 2 epochs on 1.5 M samples of thinking (math, shows, reasoning) and non-reasoning (creative writing, roleplay, easy concern answering) information. Reasoning data was produced by “expert designs”. Non-reasoning data was generated by DeepSeek-V2.5 and inspected by humans. – The “professional models” were trained by beginning with an unspecified base design, then SFT on both information, and artificial information produced by an internal DeepSeek-R1 design. The system prompt asked the R1 to show and confirm throughout thinking. Then the professional models were RL utilizing an unspecified benefit function.
– Each professional model was trained to produce simply artificial reasoning data in one particular domain (mathematics, shows, reasoning).
– Expert designs were utilized, instead of R1 itself, because the output from R1 itself suffered “overthinking, bad formatting, and extreme length”.

4. Model-based benefit designs were made by starting with a SFT checkpoint of V3, then finetuning on human choice data consisting of both final reward and chain-of-thought resulting in the final benefit. The benefit model produced benefit signals for both concerns with objective however free-form answers, and concerns without unbiased responses (such as innovative writing).
5. A SFT checkpoint of V3 was trained by GRPO utilizing both reward designs and rule-based reward. The rule-based reward was computed for mathematics issues with a last response (put in a box), and for shows issues by unit tests. This produced DeepSeek-V3.

The DeepSeek team carried out substantial low-level engineering to accomplish effectiveness. They utilized mixed-precision arithmetic. Much of the forward pass was carried out in 8-bit floating point numbers (5E2M: 5-bit exponent and 2-bit mantissa) rather than the standard 32-bit, needing unique GEMM routines to build up properly. They used a custom-made 12-bit float (E5M6) for just the inputs to the linear layers after the attention modules. Optimizer states remained in 16-bit (BF16). They minimized the interaction latency by overlapping thoroughly calculation and interaction, such as committing 20 streaming multiprocessors out of 132 per H800 for only inter-GPU interaction. They decreased communication by rearranging (every 10 minutes) the exact maker each professional was on in order to prevent specific machines being queried more typically than the others, adding auxiliary load-balancing losses to the training loss function, and other load-balancing strategies. [37]

After training, it was deployed on H800 clusters. The H800 cards within a cluster are linked by NVLink, and the clusters are linked by InfiniBand. [37]

Benchmark tests reveal that DeepSeek-V3 outshined Llama 3.1 and Qwen 2.5 whilst matching GPT-4o and Claude 3.5 Sonnet. [18] [39] [40] [41]

R1

On 20 November 2024, DeepSeek-R1-Lite-Preview ended up being available through DeepSeek’s API, along with through a chat user interface after visiting. [42] [43] [note 3] It was trained for logical inference, mathematical reasoning, and real-time problem-solving. DeepSeek declared that it went beyond efficiency of OpenAI o1 on benchmarks such as American Invitational Mathematics Examination (AIME) and MATH. [44] However, The Wall Street Journal mentioned when it used 15 issues from the 2024 edition of AIME, the o1 model reached an option quicker than DeepSeek-R1-Lite-Preview. [45]

On 20 January 2025, DeepSeek released DeepSeek-R1 and DeepSeek-R1-Zero. [46] Both were initialized from DeepSeek-V3-Base, and share its architecture. The business likewise released some “DeepSeek-R1-Distill” models, which are not initialized on V3-Base, however rather are initialized from other pretrained open-weight designs, consisting of LLaMA and Qwen, then fine-tuned on artificial data produced by R1. [47]

A discussion between User and Assistant. The user asks a concern, and the Assistant fixes it. The assistant first thinks of the reasoning procedure in the mind and then offers the user with the response. The thinking process and response are enclosed within and tags, respectively, i.e., reasoning process here respond to here. User:. Assistant:

DeepSeek-R1-Zero was trained specifically using GRPO RL without SFT. Unlike previous versions, they utilized no model-based reward. All reward functions were rule-based, “primarily” of two types (other types were not defined): accuracy rewards and format rewards. Accuracy benefit was checking whether a boxed answer is appropriate (for math) or whether a code passes tests (for shows). Format benefit was examining whether the design puts its thinking trace within … [47]

As R1-Zero has concerns with readability and mixing languages, R1 was trained to attend to these issues and more enhance reasoning: [47]

1. SFT DeepSeek-V3-Base on “thousands” of “cold-start” information all with the basic format of|special_token|| special_token|summary >.
2. Apply the very same RL process as R1-Zero, however likewise with a “language consistency reward” to motivate it to react monolingually. This produced an internal design not launched.
3. Synthesize 600K reasoning information from the internal design, with rejection sampling (i.e. if the created reasoning had a wrong final answer, then it is removed). Synthesize 200K non-reasoning information (writing, factual QA, self-cognition, translation) utilizing DeepSeek-V3.
4. SFT DeepSeek-V3-Base on the 800K artificial data for 2 epochs.
5. GRPO RL with rule-based benefit (for thinking jobs) and model-based reward (for non-reasoning tasks, helpfulness, and harmlessness). This produced DeepSeek-R1.

Distilled models were trained by SFT on 800K information synthesized from DeepSeek-R1, in a similar method as step 3 above. They were not trained with RL. [47]

Assessment and responses

DeepSeek released its AI Assistant, which utilizes the V3 model as a chatbot app for Apple IOS and Android. By 27 January 2025 the app had actually gone beyond ChatGPT as the highest-rated free app on the iOS App Store in the United States; its chatbot supposedly responds to concerns, solves reasoning issues and writes computer system programs on par with other chatbots on the marketplace, according to benchmark tests used by American AI companies. [3]

DeepSeek-V3 uses significantly less resources compared to its peers; for example, whereas the world’s leading AI business train their chatbots with supercomputers utilizing as numerous as 16,000 graphics processing units (GPUs), if not more, DeepSeek declares to require only about 2,000 GPUs, namely the H800 series chip from Nvidia. [37] It was trained in around 55 days at a cost of US$ 5.58 million, [37] which is approximately one tenth of what United States tech giant Meta invested building its newest AI technology. [3]

DeepSeek’s competitive performance at reasonably very little expense has actually been acknowledged as possibly challenging the international supremacy of American AI designs. [48] Various publications and news media, such as The Hill and The Guardian, described the release of its chatbot as a “Sputnik moment” for American AI. [49] [50] The efficiency of its R1 model was reportedly “on par with” one of OpenAI’s newest models when used for tasks such as mathematics, coding, and natural language reasoning; [51] echoing other analysts, American Silicon Valley investor Marc Andreessen also described R1 as “AI’s Sputnik minute”. [51]

DeepSeek’s founder, Liang Wenfeng has actually been compared to Open AI CEO Sam Altman, with CNN calling him the Sam Altman of China and an evangelist for AI. [52] Chinese state media widely applauded DeepSeek as a nationwide asset. [53] [54] On 20 January 2025, China’s Premier Li Qiang invited Liang Wenfeng to his symposium with experts and asked him to supply opinions and tips on a draft for comments of the yearly 2024 government work report. [55]

DeepSeek’s optimization of restricted resources has actually highlighted possible limits of United States sanctions on China’s AI development, that include export restrictions on advanced AI chips to China [18] [56] The success of the business’s AI designs as a result “triggered market chaos” [57] and caused shares in significant global innovation companies to plunge on 27 January 2025: Nvidia’s stock fell by as much as 17-18%, [58] as did the stock of rival Broadcom. Other tech firms likewise sank, including Microsoft (down 2.5%), Google’s owner Alphabet (down over 4%), and Dutch chip equipment maker ASML (down over 7%). [51] A worldwide selloff of innovation stocks on Nasdaq, prompted by the release of the R1 design, had actually led to record losses of about $593 billion in the market capitalizations of AI and hardware companies; [59] by 28 January 2025, a total of $1 trillion of worth was rubbed out American stocks. [50]

Leading figures in the American AI sector had combined reactions to DeepSeek’s success and efficiency. [60] Microsoft CEO Satya Nadella and OpenAI CEO Sam Altman-whose companies are associated with the United States government-backed “Stargate Project” to develop American AI infrastructure-both called DeepSeek “incredibly outstanding”. [61] [62] American President Donald Trump, who announced The Stargate Project, called DeepSeek a wake-up call [63] and a favorable advancement. [64] [50] [51] [65] Other leaders in the field, consisting of Scale AI CEO Alexandr Wang, Anthropic cofounder and CEO Dario Amodei, and Elon Musk expressed uncertainty of the app’s efficiency or of the sustainability of its success. [60] [66] [67] Various companies, including Amazon Web Services, Toyota, and Stripe, are seeking to utilize the model in their program. [68]

On 27 January 2025, DeepSeek limited its brand-new user registration to telephone number from mainland China, e-mail addresses, or Google account logins, following a “large-scale” cyberattack disrupted the proper performance of its servers. [69] [70]

Some sources have observed that the main application programming user interface (API) variation of R1, which runs from servers located in China, uses censorship mechanisms for subjects that are considered politically sensitive for the federal government of China. For instance, the model refuses to answer questions about the 1989 Tiananmen Square demonstrations and massacre, persecution of Uyghurs, contrasts in between Xi Jinping and Winnie the Pooh, or human rights in China. [71] [72] [73] The AI might initially create an answer, however then erases it quickly afterwards and changes it with a message such as: “Sorry, that’s beyond my current scope. Let’s discuss something else.” [72] The incorporated censorship systems and restrictions can just be gotten rid of to a restricted degree in the open-source version of the R1 design. If the “core socialist worths” specified by the Chinese Internet regulative authorities are discussed, or the political status of Taiwan is raised, discussions are terminated. [74] When evaluated by NBC News, DeepSeek’s R1 described Taiwan as “an inalienable part of China’s area,” and stated: “We firmly oppose any form of ‘Taiwan independence’ separatist activities and are dedicated to accomplishing the total reunification of the motherland through peaceful means.” [75] In January 2025, Western scientists were able to deceive DeepSeek into offering specific responses to some of these topics by requesting in its answer to switch particular letters for similar-looking numbers. [73]

Security and privacy

Some specialists fear that the government of China might use the AI system for foreign influence operations, spreading disinformation, monitoring and the advancement of cyberweapons. [76] [77] [78] DeepSeek’s privacy terms and conditions state “We store the information we collect in secure servers found in the People’s Republic of China … We might collect your text or audio input, timely, uploaded files, feedback, chat history, or other content that you provide to our design and Services”. Although the data storage and collection policy follows ChatGPT’s privacy policy, [79] a Wired short article reports this as security concerns. [80] In response, the Italian information defense authority is seeking additional details on DeepSeek’s collection and use of personal data, and the United States National Security Council revealed that it had actually started a nationwide security evaluation. [81] [82] Taiwan’s federal government prohibited using DeepSeek at federal government ministries on security premises and South Korea’s Personal Information Protection Commission opened a query into DeepSeek’s use of individual details. [83]

Artificial intelligence market in China.

Notes

^ a b c The variety of heads does not equal the number of KV heads, due to GQA.
^ Inexplicably, the model called DeepSeek-Coder-V2 Chat in the paper was launched as DeepSeek-Coder-V2-Instruct in HuggingFace.
^ At that time, the R1-Lite-Preview required selecting “Deep Think allowed”, and every user might use it just 50 times a day.
References

^ Gibney, Elizabeth (23 January 2025). “China’s cheap, open AI design DeepSeek thrills scientists”. Nature. doi:10.1038/ d41586-025-00229-6. ISSN 1476-4687. PMID 39849139.
^ a b Vincent, James (28 January 2025). “The DeepSeek panic reveals an AI world prepared to blow”. The Guardian.
^ a b c d e f g Metz, Cade; Tobin, Meaghan (23 January 2025). “How Chinese A.I. Start-Up DeepSeek Is Taking On Silicon Valley Giants”. The New York Times. ISSN 0362-4331. Retrieved 27 January 2025.
^ Cosgrove, Emma (27 January 2025). “DeepSeek’s cheaper designs and weaker chips bring into question trillions in AI facilities costs”. Business Insider.
^ Mallick, Subhrojit (16 January 2024). “Biden admin’s cap on GPU exports may hit India’s AI aspirations”. The Economic Times. Retrieved 29 January 2025.
^ Saran, Cliff (10 December 2024). “Nvidia examination signals expanding of US and China chip war|Computer Weekly”. Computer Weekly. Retrieved 27 January 2025.
^ Sherman, Natalie (9 December 2024). “Nvidia targeted by China in new chip war probe”. BBC. Retrieved 27 January 2025.
^ a b c Metz, Cade (27 January 2025). “What is DeepSeek? And How Is It Upending A.I.?”. The New York City Times. ISSN 0362-4331. Retrieved 27 January 2025.
^ Field, Hayden (27 January 2025). “China’s DeepSeek AI dethrones ChatGPT on App Store: Here’s what you ought to understand”. CNBC.
^ Picchi, Aimee (27 January 2025). “What is DeepSeek, and why is it causing Nvidia and other stocks to slump?”. CBS News.
^ Zahn, Max (27 January 2025). “Nvidia, Microsoft shares tumble as China-based AI app DeepSeek hammers tech giants”. ABC News. Retrieved 27 January 2025.
^ Roose, Kevin (28 January 2025). “Why DeepSeek Could Change What Silicon Valley Believe About A.I.” The New York Times. ISSN 0362-4331. Retrieved 28 January 2025.
^ a b Romero, Luis E. (28 January 2025). “ChatGPT, DeepSeek, Or Llama? Meta’s LeCun Says Open-Source Is The Key”. Forbes.
^ Chen, Caiwei (24 January 2025). “How a top Chinese AI design overcame US sanctions”. MIT Technology Review. Archived from the original on 25 January 2025. Retrieved 25 January 2025.
^ a b c d Ottinger, Lily (9 December 2024). “Deepseek: From Hedge Fund to Frontier Model Maker”. ChinaTalk. Archived from the original on 28 December 2024. Retrieved 28 December 2024.
^ Leswing, Kif (23 February 2023). “Meet the $10,000 Nvidia chip powering the race for A.I.” CNBC. Retrieved 30 January 2025.
^ Yu, Xu (17 April 2023).” [Exclusive] Chinese Quant Hedge Fund High-Flyer Won’t Use AGI to Trade Stocks, MD Says”. Yicai Global. Archived from the initial on 31 December 2023. Retrieved 28 December 2024.
^ a b c d e Jiang, Ben; Perezi, Bien (1 January 2025). “Meet DeepSeek: the Chinese start-up that is changing how AI designs are trained”. South China Morning Post. Archived from the initial on 22 January 2025. Retrieved 1 January 2025.
^ a b McMorrow, Ryan; Olcott, Eleanor (9 June 2024). “The Chinese quant fund-turned-AI pioneer”. Financial Times. Archived from the original on 17 July 2024. Retrieved 28 December 2024.
^ a b Schneider, Jordan (27 November 2024). “Deepseek: The Quiet Giant Leading China’s AI Race”. ChinaTalk. Retrieved 28 December 2024.
^ “DeepSeek-Coder/LICENSE-MODEL at primary · deepseek-ai/DeepSeek-Coder”. GitHub. Archived from the original on 22 January 2025. Retrieved 24 January 2025.
^ a b c Guo, Daya; Zhu, Qihao; Yang, Dejian; Xie, Zhenda; Dong, Kai; Zhang, Wentao; Chen, Guanting; Bi, Xiao; Wu, Y. (26 January 2024), DeepSeek-Coder: When the Large Language Model Meets Programming – The Rise of Code Intelligence, arXiv:2401.14196.
^ “DeepSeek Coder”. deepseekcoder.github.io. Retrieved 27 January 2025.
^ deepseek-ai/DeepSeek-Coder, DeepSeek, 27 January 2025, recovered 27 January 2025.
^ “deepseek-ai/deepseek-coder -5.7 bmqa-base · Hugging Face”. huggingface.co. Retrieved 27 January 2025.
^ a b c d DeepSeek-AI; Bi, Xiao; Chen, Deli; Chen, Guanting; Chen, Shanhuang; Dai, Damai; Deng, Chengqi; Ding, Honghui; Dong, Kai (5 January 2024), DeepSeek LLM: Scaling Open-Source Language Models with Longtermism, arXiv:2401.02954.
^ deepseek-ai/DeepSeek-LLM, DeepSeek, 27 January 2025, retrieved 27 January 2025.
^ a b Dai, Damai; Deng, Chengqi; Zhao, Chenggang; Xu, R. X.; Gao, Huazuo; Chen, Deli; Li, Jiashi; Zeng, Wangding; Yu, Xingkai (11 January 2024), DeepSeekMoE: Towards Ultimate Expert Specialization in Mixture-of-Experts Language Models, arXiv:2401.06066.
^ Shao, Zhihong; Wang, Peiyi; Zhu, Qihao; Xu, Runxin; Song, Junxiao; Bi, Xiao; Zhang, Haowei; Zhang, Mingchuan; Li, Y. K. (27 April 2024), DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models, arXiv:2402.03300.
^ Wang, Peiyi; Li, Lei; Shao, Zhihong; Xu, R. X.; Dai, Damai; Li, Yifei; Chen, Deli; Wu, Y.; Sui, Zhifang (19 February 2024), Math-Shepherd: Verify and Reinforce LLMs Step-by-step without Human Annotations, arXiv:2312.08935. ^ a b c d DeepSeek-AI; Liu, Aixin; Feng, Bei; Wang, Bin; Wang, Bingxuan; Liu, Bo; Zhao, Chenggang; Dengr, Chengqi; Ruan, Chong (19 June 2024), DeepSeek-V2: A Strong, Economical, and Efficient Mixture-of-Experts Language Model, arXiv:2405.04434.
^ a b Peng, Bowen; Quesnelle, Jeffrey; Fan, Honglu; Shippole, Enrico (1 November 2023), YaRN: Efficient Context Window Extension of Large Language Models, arXiv:2309.00071.
^ “config.json · deepseek-ai/DeepSeek-V 2-Lite at primary”. huggingface.co. 15 May 2024. Retrieved 28 January 2025.
^ “config.json · deepseek-ai/DeepSeek-V 2 at primary”. huggingface.co. 6 May 2024. Retrieved 28 January 2025.
^ DeepSeek-AI; Zhu, Qihao; Guo, Daya; Shao, Zhihong; Yang, Dejian; Wang, Peiyi; Xu, Runxin; Wu, Y.; Li, Yukun (17 June 2024), DeepSeek-Coder-V2: Breaking the Barrier of Closed-Source Models in Code Intelligence, arXiv:2406.11931.
^ “deepseek-ai/DeepSeek-V 2.5 · Hugging Face”. huggingface.co. 3 January 2025. Retrieved 28 January 2025.
^ a b c d e f g DeepSeek-AI; Liu, Aixin; Feng, Bei; Xue, Bing; Wang, Bingxuan; Wu, Bochao; Lu, Chengda; Zhao, Chenggang; Deng, Chengqi (27 December 2024), DeepSeek-V3 Technical Report, arXiv:2412.19437.
^ “config.json · deepseek-ai/DeepSeek-V 3 at main”. huggingface.co. 26 December 2024. Retrieved 28 January 2025.
^ Jiang, Ben (27 December 2024). “Chinese start-up DeepSeek’s brand-new AI model exceeds Meta, OpenAI products”. South China Morning Post. Archived from the original on 27 December 2024. Retrieved 28 December 2024.
^ Sharma, Shubham (26 December 2024). “DeepSeek-V3, ultra-large open-source AI, outshines Llama and Qwen on launch”. VentureBeat. Archived from the original on 27 December 2024. Retrieved 28 December 2024.
^ Wiggers, Kyle (26 December 2024). “DeepSeek’s brand-new AI model seems among the finest ‘open’ oppositions yet”. TechCrunch. Archived from the initial on 2 January 2025. Retrieved 31 December 2024.
^ “Deepseek Log in page”. DeepSeek. Retrieved 30 January 2025.
^ “News|DeepSeek-R1-Lite Release 2024/11/20: DeepSeek-R1-Lite-Preview is now live: releasing supercharged reasoning power!”. DeepSeek API Docs. Archived from the original on 20 November 2024. Retrieved 28 January 2025.
^ Franzen, Carl (20 November 2024). “DeepSeek’s first reasoning model R1-Lite-Preview turns heads, beating OpenAI o1 performance”. VentureBeat. Archived from the initial on 22 November 2024. Retrieved 28 December 2024.
^ Huang, Raffaele (24 December 2024). “Don’t Look Now, however China’s AI Is Catching Up Fast”. The Wall Street Journal. Archived from the original on 27 December 2024. Retrieved 28 December 2024.
^ “Release DeepSeek-R1 · deepseek-ai/DeepSeek-R1@23807ce”. GitHub. Archived from the initial on 21 January 2025. Retrieved 21 January 2025.
^ a b c d DeepSeek-AI; Guo, Daya; Yang, Dejian; Zhang, Haowei; Song, Junxiao; Zhang, Ruoyu; Xu, Runxin; Zhu, Qihao; Ma, Shirong (22 January 2025), DeepSeek-R1: Incentivizing Reasoning Capability in LLMs through Reinforcement Learning, arXiv:2501.12948.
^ “Chinese AI start-up DeepSeek overtakes ChatGPT on Apple App Store”. Reuters. 27 January 2025. Retrieved 27 January 2025.
^ Wade, David (6 December 2024). “American AI has actually reached its Sputnik minute”. The Hill. Archived from the initial on 8 December 2024. Retrieved 25 January 2025.
^ a b c Milmo, Dan; Hawkins, Amy; Booth, Robert; Kollewe, Julia (28 January 2025). “‘ Sputnik minute’: $1tn rubbed out US stocks after Chinese firm reveals AI chatbot” – via The Guardian.
^ a b c d Hoskins, Peter; Rahman-Jones, Imran (27 January 2025). “Nvidia shares sink as Chinese AI app spooks markets”. BBC. Retrieved 28 January 2025.
^ Goldman, David (27 January 2025). “What is DeepSeek, the Chinese AI start-up that shook the tech world?|CNN Business”. CNN. Retrieved 29 January 2025.
^ “DeepSeek positions a difficulty to Beijing as much as to Silicon Valley”. The Economist. 29 January 2025. ISSN 0013-0613. Retrieved 31 January 2025.
^ Paul, Katie; Nellis, Stephen (30 January 2025). “Chinese state-linked accounts hyped DeepSeek AI launch ahead of US stock rout, Graphika says”. Reuters. Retrieved 30 January 2025.
^ 澎湃新闻 (22 January 2025). “量化巨头幻方创始人梁文锋参加总理座谈会并发言 ， 他还创办了” AI界拼多多””. finance.sina.com.cn. Retrieved 31 January 2025.
^ Shilov, Anton (27 December 2024). “Chinese AI company’s AI design breakthrough highlights limitations of US sanctions”. Tom’s Hardware. Archived from the initial on 28 December 2024. Retrieved 28 December 2024.
^ “DeepSeek updates – Chinese AI chatbot stimulates US market chaos, wiping $500bn off Nvidia”. BBC News. Retrieved 27 January 2025.
^ Nazareth, Rita (26 January 2025). “Stock Rout Gets Ugly as Nvidia Extends Loss to 17%: Markets Wrap”. Bloomberg. Retrieved 27 January 2025.
^ Carew, Sinéad; Cooper, Amanda; Banerjee, Ankur (27 January 2025). “DeepSeek stimulates worldwide AI selloff, Nvidia losses about $593 billion of worth”. Reuters.
^ a b Sherry, Ben (28 January 2025). “DeepSeek, Calling It ‘Impressive’ but Staying Skeptical”. Inc. Retrieved 29 January 2025.
^ Okemwa, Kevin (28 January 2025). “Microsoft CEO Satya Nadella touts DeepSeek’s open-source AI as “extremely outstanding”: “We should take the developments out of China very, extremely seriously””. Windows Central. Retrieved 28 January 2025.
^ Nazzaro, Miranda (28 January 2025). “OpenAI’s Sam Altman calls DeepSeek design ‘remarkable'”. The Hill. Retrieved 28 January 2025.
^ Dou, Eva; Gregg, Aaron; Zakrzewski, Cat; Tiku, Nitasha; Najmabadi, Shannon (28 January 2025). “Trump calls China’s DeepSeek AI app a ‘wake-up call’ after tech stocks slide”. The Washington Post. Retrieved 28 January 2025.
^ Habeshian, Sareen (28 January 2025). “Johnson bashes China on AI, Trump calls DeepSeek development “favorable””. Axios.
^ Karaian, Jason; Rennison, Joe (27 January 2025). “China’s A.I. Advances Spook Big Tech Investors on Wall Street” – via NYTimes.com.
^ Sharma, Manoj (6 January 2025). “Musk dismisses, Altman praises: What leaders say on DeepSeek’s disturbance”. Fortune India. Retrieved 28 January 2025.
^ “Elon Musk ‘questions’ DeepSeek’s claims, suggests massive Nvidia GPU infrastructure”. Financialexpress. 28 January 2025. Retrieved 28 January 2025.
^ Kim, Eugene. “Big AWS clients, consisting of Stripe and Toyota, are pestering the cloud giant for access to DeepSeek AI models”. Business Insider.
^ Kerr, Dara (27 January 2025). “DeepSeek hit with ‘large-scale’ cyber-attack after AI chatbot tops app stores”. The Guardian. Retrieved 28 January 2025.
^ Tweedie, Steven; Altchek, Ana. “DeepSeek temporarily restricted new sign-ups, pointing out ‘large-scale destructive attacks'”. Business Insider.
^ Field, Matthew; Titcomb, James (27 January 2025). “Chinese AI has triggered a $1 trillion panic – and it doesn’t appreciate free speech”. The Daily Telegraph. ISSN 0307-1235. Retrieved 27 January 2025.
^ a b Steinschaden, Jakob (27 January 2025). “DeepSeek: This is what live censorship looks like in the Chinese AI chatbot”. Trending Topics. Retrieved 27 January 2025.
^ a b Lu, Donna (28 January 2025). “We attempted out DeepSeek. It worked well, until we asked it about Tiananmen Square and Taiwan”. The Guardian. ISSN 0261-3077. Retrieved 30 January 2025.
^ “The Guardian view on a worldwide AI race: geopolitics, innovation and the rise of chaos”. The Guardian. 26 January 2025. ISSN 0261-3077. Retrieved 27 January 2025.
^ Yang, Angela; Cui, Jasmine (27 January 2025). “Chinese AI DeepSeek jolts Silicon Valley, giving the AI race its ‘Sputnik moment'”. NBC News. Retrieved 27 January 2025.
^ Kimery, Anthony (26 January 2025). “China’s DeepSeek AI positions formidable cyber, information privacy threats”. Biometric Update. Retrieved 27 January 2025.
^ Booth, Robert; Milmo, Dan (28 January 2025). “Experts prompt caution over use of Chinese AI DeepSeek”. The Guardian. ISSN 0261-3077. Retrieved 28 January 2025.
^ Hornby, Rael (28 January 2025). “DeepSeek’s success has painted a substantial TikTok-shaped target on its back”. LaptopMag. Retrieved 28 January 2025.
^ “Privacy policy”. Open AI. Retrieved 28 January 2025.
^ Burgess, Matt; Newman, Lily Hay (27 January 2025). “DeepSeek’s Popular AI App Is Explicitly Sending US Data to China”. Wired. ISSN 1059-1028. Retrieved 28 January 2025.
^ “Italy regulator inquires from DeepSeek on information defense”. Reuters. 28 January 2025. Retrieved 28 January 2025.
^ Shalal, Andrea; Shepardson, David (28 January 2025). “White House examines result of China AI app DeepSeek on national security, authorities says”. Reuters. Retrieved 28 January 2025.