Topzle Topzle

DeepSeek

Updated: 12/10/2025, 4:53:32 PM Wikipedia source

Hangzhou DeepSeek Artificial Intelligence Basic Technology Research Co., Ltd., doing business as DeepSeek, is a Chinese artificial intelligence (AI) company that develops large language models (LLMs). Based in Hangzhou, Zhejiang, Deepseek is owned and funded by the Chinese hedge fund High-Flyer. DeepSeek was founded in July 2023 by Liang Wenfeng, the co-founder of High-Flyer, who also serves as the CEO for both of the companies. The company launched an eponymous chatbot alongside its DeepSeek-R1 model in January 2025. Released under the MIT License, DeepSeek-R1 provides responses comparable to other contemporary large language models, such as OpenAI's GPT-4 and o1. Its training cost was reported to be significantly lower than other LLMs. The company claims that it trained its V3 model for US$6 million—far less than the US$100 million cost for OpenAI's GPT-4 in 2023—and using approximately one-tenth the computing power consumed by Meta's comparable model, Llama 3.1. DeepSeek's success against larger and more established rivals has been described as "upending AI". DeepSeek's models are described as "open weight," meaning the exact parameters are openly shared, although certain usage conditions differ from typical open-source software. The company reportedly recruits AI researchers from top Chinese universities and also hires from outside traditional computer science fields to broaden its models' knowledge and capabilities. DeepSeek significantly reduced training expenses for their R1 model by incorporating techniques such as mixture of experts (MoE) layers. The company also trained its models during ongoing trade restrictions on AI chip exports to China, using weaker AI chips intended for export and employing fewer units overall. Observers say this breakthrough sent "shock waves" through the industry which were described as triggering a "Sputnik moment" for the US in the field of artificial intelligence, particularly due to its open-source, cost-effective, and high-performing AI models. This threatened established AI hardware leaders such as Nvidia; Nvidia's share price dropped sharply, losing US$600 billion in market value, the largest single-company decline in U.S. stock market history.

Infobox

Native name
杭州深度求索人工智能基础技术研究有限公司
Company type
Private
Industry
Information technologyArtificial intelligence
Founded
17 July 2023; 2 years ago (2023-07-17)
Founder
mw- Liang Wenfeng
Headquarters
Hangzhou, Zhejiang, China
Key people
Liang Wenfeng (CEO)
Owner
High-Flyer
Number of employees
160 (2025)
Website
deepseek.com

Tables

Major versions of DeepSeek models. SFT stands for supervised finetuning. · Development and release history
DeepSeek Coder
DeepSeek Coder
Major versions
DeepSeek Coder
Release date
November 2, 2023
Status
Discontinued
Major variants
Base (pretrained); Instruct (with instruction-finetuned)
Remarks
The architecture is essentially the same as Llama.
DeepSeek-LLM
DeepSeek-LLM
Major versions
DeepSeek-LLM
Release date
November 29, 2023
Status
Discontinued
Major variants
Base; Chat (with SFT)
DeepSeek-MoE
DeepSeek-MoE
Major versions
DeepSeek-MoE
Release date
January 9, 2024
Status
Discontinued
Major variants
Base; Chat
Remarks
Developed a variant of mixture of experts (MoE).
DeepSeek-Math
DeepSeek-Math
Major versions
DeepSeek-Math
Release date
April 2024
Status
Discontinued
Major variants
Base
Remarks
Initialized with DS-Coder-Base-v1.5
Instruct (with SFT)
Instruct (with SFT)
Major versions
Instruct (with SFT)
RL (using a process reward model)
RL (using a process reward model)
Major versions
RL (using a process reward model)
Release date
Developed Group Relative Policy Optimization (GRPO), a variant of Proximal Policy Optimization (PPO).
DeepSeek V2
DeepSeek V2
Major versions
DeepSeek V2
Release date
May 2024
Status
Discontinued
Major variants
DeepSeek-V2, DeepSeek-V2-Chat DeepSeek-V2-Lite, DeepSeek-V2-Lite-Chat DeepSeek-Coder-V2 DeepSeek-V2.5
Remarks
Developed multi-head latent attention (MLA). Also used mixture of experts (MoE). Implemented KV caching.
DeepSeek V3
DeepSeek V3
Major versions
DeepSeek V3
Release date
December 2024
Status
Active
Major variants
DeepSeek-V3-BaseDeepSeek-V3 (a chat model)
Remarks
The architecture is essentially the same as V2. Updated on 2025-03-24.
DeepSeek-Prover-V2
DeepSeek-Prover-V2
Major versions
DeepSeek-Prover-V2
Release date
May 1, 2025
Status
Active
Major variants
DeepSeek-Prover-V2-671BDeepSeek-Prover-V2-7B
DeepSeek VL2
DeepSeek VL2
Major versions
DeepSeek VL2
Release date
December 13, 2024
Status
Active
DeepSeek R1
DeepSeek R1
Major versions
DeepSeek R1
Release date
November 20, 2024
Status
Active
Major variants
DeepSeek-R1-Lite-Preview
Remarks
Only accessed through API and a chat interface.
January 20, 2025
January 20, 2025
Major versions
January 20, 2025
Release date
Active
Status
DeepSeek-R1 DeepSeek-R1-Zero
Major variants
Initialized from DeepSeek-V3-Base and sharing the V3 architecture.
Distilled models
Distilled models
Major versions
Distilled models
Release date
Initialized from other models, such as Llama, Qwen, etc. Distilled from data synthesized by R1 and R1-Zero.
May 28, 2025
May 28, 2025
Major versions
May 28, 2025
Release date
Active
Status
DeepSeek-R1-0528
DeepSeek V3.1
DeepSeek V3.1
Major versions
DeepSeek V3.1
Release date
August 21, 2025
Status
Active
Major variants
DeepSeek-V3.1-BaseDeepSeek-V3.1 (a chat model)
Remarks
Hybrid architecture (thinking and non-thinking modes available). Trained on over 800B additional tokens on top of V3.
September 22, 2025
September 22, 2025
Major versions
September 22, 2025
Release date
Active
Status
DeepSeek-V3.1-Terminus
Major variants
Reducing instances of mixed Chinese-English text and occasional abnormal characters on top of V3.1.
DeepSeekMath-V2
DeepSeekMath-V2
Major versions
DeepSeekMath-V2
Release date
November 27, 2025
Status
Active
Major versions
Release date
Status
Major variants
Remarks
DeepSeek Coder
November 2, 2023
Discontinued
Base (pretrained); Instruct (with instruction-finetuned)
The architecture is essentially the same as Llama.
DeepSeek-LLM
November 29, 2023
Discontinued
Base; Chat (with SFT)
DeepSeek-MoE
January 9, 2024
Discontinued
Base; Chat
Developed a variant of mixture of experts (MoE).
DeepSeek-Math
April 2024
Discontinued
Base
Initialized with DS-Coder-Base-v1.5
Instruct (with SFT)
RL (using a process reward model)
Developed Group Relative Policy Optimization (GRPO), a variant of Proximal Policy Optimization (PPO).
DeepSeek V2
May 2024
Discontinued
DeepSeek-V2, DeepSeek-V2-Chat DeepSeek-V2-Lite, DeepSeek-V2-Lite-Chat DeepSeek-Coder-V2 DeepSeek-V2.5
Developed multi-head latent attention (MLA). Also used mixture of experts (MoE). Implemented KV caching.
DeepSeek V3
December 2024
Active
DeepSeek-V3-BaseDeepSeek-V3 (a chat model)
The architecture is essentially the same as V2. Updated on 2025-03-24.
DeepSeek-Prover-V2
May 1, 2025
Active
DeepSeek-Prover-V2-671BDeepSeek-Prover-V2-7B
DeepSeek VL2
December 13, 2024
Active
DeepSeek R1
November 20, 2024
Active
DeepSeek-R1-Lite-Preview
Only accessed through API and a chat interface.
January 20, 2025
Active
DeepSeek-R1 DeepSeek-R1-Zero
Initialized from DeepSeek-V3-Base and sharing the V3 architecture.
Distilled models
Initialized from other models, such as Llama, Qwen, etc. Distilled from data synthesized by R1 and R1-Zero.
May 28, 2025
Active
DeepSeek-R1-0528
DeepSeek V3.1
August 21, 2025
Active
DeepSeek-V3.1-BaseDeepSeek-V3.1 (a chat model)
Hybrid architecture (thinking and non-thinking modes available). Trained on over 800B additional tokens on top of V3.
September 22, 2025
Active
DeepSeek-V3.1-Terminus
Reducing instances of mixed Chinese-English text and occasional abnormal characters on top of V3.1.
DeepSeekMath-V2
November 27, 2025
Active
DeepSeek Coder properties[64]: Table 2 [67] · Overview of models › DeepSeek Coder
1.3B
1.3B
Params.
1.3B
# Layers
24
Model dim.
2048
Intermediate dim.
5504
# Heads
16
# Kv-heads
16
5.7B
5.7B
Params.
5.7B
# Layers
32
Model dim.
4096
Intermediate dim.
11008
# Heads
32
# Kv-heads
1
6.7B
6.7B
Params.
6.7B
# Layers
32
Model dim.
4096
Intermediate dim.
11008
# Heads
32
# Kv-heads
32
33B
33B
Params.
33B
# Layers
62
Model dim.
7168
Intermediate dim.
19200
# Heads
56
# Kv-heads
7
Params.
# Layers
Model dim.
Intermediate dim.
# Heads
# Kv-heads
1.3B
24
2048
5504
16
16
5.7B
32
4096
11008
32
1
6.7B
32
4096
11008
32
32
33B
62
7168
19200
56
7
DeepSeek LLM properties[36]: Table 2 · Overview of models › DeepSeek-LLM
7B
7B
Params.
7B
# Layers
30
Model dim.
4096
Intermediate dim.
11008
# Heads
32
# Kv-heads
32
67B
67B
Params.
67B
# Layers
95
Model dim.
8192
Intermediate dim.
22016
# Heads
64
# Kv-heads
8
Params.
# Layers
Model dim.
Intermediate dim.
# Heads
# Kv-heads
7B
30
4096
11008
32
32
67B
95
8192
22016
64
8
DeepSeek V2 properties[70]: Section 3.1.2, Appendix B [72][73] · Overview of models › V2
V2-Lite
V2-Lite
Name
V2-Lite
Params.
15.7B
Active params
2.4B
# Layers
27
Context length
32K
# Shared experts
2
# Routed experts
64
V2
V2
Name
V2
Params.
236B
Active params
21B
# Layers
60
Context length
128K
# Shared experts
2
# Routed experts
160
Name
Params.
Active params
# Layers
Context length
# Shared experts
# Routed experts
V2-Lite
15.7B
2.4B
27
32K
2
64
V2
236B
21B
60
128K
2
160
DeepSeek V3 properties[30]: Section 4.2 [75] · Overview of models › V3
V3
V3
Name
V3
Params.
671B
Active params
37B
# Layers
61
Context length
128K
# Shared experts
1
# Routed experts
256
Name
Params.
Active params
# Layers
Context length
# Shared experts
# Routed experts
V3
671B
37B
61
128K
1
256
Total cost of training the DeepSeek-V3 model[30]: Table 1 · Overview of models › V3
Pre-training
Pre-training
Stage
Pre-training
Cost (in one thousand GPU hours)
2,664
Cost (in one million US$)
5.328
Context extension
Context extension
Stage
Context extension
Cost (in one thousand GPU hours)
119
Cost (in one million US$)
0.24
Fine-tuning
Fine-tuning
Stage
Fine-tuning
Cost (in one thousand GPU hours)
5
Cost (in one million US$)
0.01
Total
Total
Stage
Total
Cost (in one thousand GPU hours)
2,788
Cost (in one million US$)
5.576
Stage
Cost (in one thousand GPU hours)
Cost (in one million US$)
Pre-training
2,664
5.328
Context extension
119
0.24
Fine-tuning
5
0.01
Total
2,788
5.576

References

  1. Chinese: 杭州深度求索人工智能基础技术研究有限公司. Sometimes simply referred to in English as Hangzhou DeepSeek Artificial Intelligence.
  2. Chinese: 深度求索; pinyin: Shēndù Qiúsuǒ
  3. 宁波程信柔兆企业管理咨询合伙企业（有限合伙） and 宁波程恩企业管理咨询合伙企业（有限合伙）
  4. The number of heads does not equal the number of KV heads, due to GQA.
  5. Inexplicably, the model named DeepSeek-Coder-V2 Chat in the paper was released as DeepSeek-Coder-V2-Instruct in HuggingF
  6. At that time, the R1-Lite-Preview required selecting "Deep Think enabled", and every user could use it only 50 times a d
  7. "DeepSeek突传消息"
    https://finance.sina.com.cn/jjxw/2025-02-01/doc-inehyqcx9694053.shtml
  8. Financial Times
    https://www.ft.com/content/fb5c11bb-1d4b-465f-8283-451a19a3d425
  9. Bloomberg L.P.
    https://www.bloomberg.com/profile/company/2544189D:CH
  10. DeepSeek
    https://chat.deepseek.com/downloads/DeepSeek%20Coder%20Model%20Service%20Agreement_1019.pdf
  11. DeepSeek
    https://chat.deepseek.com/downloads/DeepSeek%20Coder%20Privacy%20Policy_1019.pdf
  12. beian.mps.gov.cn
    https://beian.mps.gov.cn/#/query/webSearch?code=33010502011812
  13. South China Morning Post
    https://www.scmp.com/tech/policy/article/3295662/beijing-meeting-puts-spotlight-chinas-new-face-ai-deepseek-founder-liang-wenfeng
  14. Reuters
    https://www.reuters.com/technology/deepseek-founder-liang-wenfeng-puts-focus-chinese-innovation-2025-01-28/
  15. The Economist
    https://www.economist.com/china/2025/02/19/behind-deepseek-lies-a-dazzling-chinese-university
  16. Nature
    https://www.nature.com/articles/d41586-025-00229-6
  17. The Guardian
    https://www.theguardian.com/commentisfree/2025/jan/28/deepseek-r1-ai-world-chinese-chatbot-tech-world-western
  18. The New York Times
    https://www.nytimes.com/2025/01/23/technology/deepseek-china-ai-chips.html
  19. Business Insider
    https://www.businessinsider.com/explaining-deepseek-chinese-models-efficiency-scaring-markets-2025-1
  20. The New York Times
    https://www.nytimes.com/2025/01/27/technology/what-is-deepseek-china-ai.html
  21. The New York Times
    https://www.nytimes.com/2025/01/28/technology/why-deepseek-could-change-what-silicon-valley-believes-about-ai.html
  22. Popular Mechanics
    https://www.popularmechanics.com/science/a63633889/deepseek-open-weight/
  23. The New York Times
    https://www.nytimes.com/2025/02/12/technology/deepseek-ai-chip-costs.html
  24. Center for Strategic and International Studies
    https://www.csis.org/analysis/deepseek-huawei-export-controls-and-future-us-china-ai-race
  25. The Guardian
    https://www.theguardian.com/technology/2025/jan/28/who-is-behind-deepseek-and-how-did-it-achieve-its-ai-sputnik-moment
  26. The New Yorker
    https://www.newyorker.com/news/the-financial-page/is-deepseek-chinas-sputnik-moment
  27. NPR
    https://www.npr.org/2025/01/28/g-s1-45061/deepseek-did-a-little-known-chinese-startup-cause-a-sputnik-moment-for-ai
  28. Liberation News – The Newspaper of the Party for Socialism and Liberation
    https://liberationnews.org/deepseek-sends-shock-waves-across-silicon-valley/
  29. Sky News
    https://news.sky.com/story/deepseek-us-tech-stocks-tumble-on-fears-of-cheaper-chinese-ai-13297788
  30. MIT Technology Review
    https://www.technologyreview.com/2025/01/24/1110526/china-deepseek-top-ai-despite-sanctions/
  31. High-Flyer
    https://www.high-flyer.cn/history/
  32. ChinaTalk
    https://www.chinatalk.media/p/deepseek-from-hedge-fund-to-frontier
  33. Financial Times
    https://www.ft.com/content/747a7b11-dcba-4aa5-8d25-403f56216d7e
  34. CNBC
    https://www.cnbc.com/2023/02/23/nvidias-a100-is-the-10000-chip-powering-the-race-for-ai-.html
  35. High-Flyer
    https://www.high-flyer.cn/blog/hf-reduce/
  36. DeepSeek-V3 Technical Report
    https://arxiv.org/abs/2412.19437
  37. SC24: International Conference for High Performance Computing, Networking, Storage and Analysis
    https://arxiv.org/abs/2408.14158
  38. Yicai
    https://www.yicai.com/news/101732215.html
  39. Yicai Global
    https://www.yicaiglobal.com/news/exclusive-chinese-quant-fund-high-flyer-will-not-use-agi-to-trade-stocks-managing-director-says
  40. South China Morning Post
    https://www.scmp.com/tech/tech-trends/article/3293050/meet-deepseek-chinese-start-changing-how-ai-models-are-trained
  41. Financial Times
    https://www.ft.com/content/357f3c68-b866-4c2e-b678-0d075051a260
  42. DeepSeek LLM: Scaling Open-Source Language Models with Longtermism
    https://arxiv.org/abs/2401.02954
  43. DeepSeekMoE: Towards Ultimate Expert Specialization in Mixture-of-Experts Language Models
    https://arxiv.org/abs/2401.06066
  44. DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models
    https://arxiv.org/abs/2402.03300
  45. DeepSeek-Coder-V2: Breaking the Barrier of Closed-Source Models in Code Intelligence
    https://arxiv.org/abs/2406.11931
  46. Hugging Face
    https://huggingface.co/deepseek-ai/DeepSeek-V2.5
  47. DeepSeek
    https://chat.deepseek.com/sign_in
  48. DeepSeek API Docs
    https://web.archive.org/web/20241120141324/https://api-docs.deepseek.com/news/news1120
  49. CNBC
    https://www.cnbc.com/2025/01/27/chinas-deepseek-ai-tops-chatgpt-app-store-what-you-should-know.html
  50. CBS News
    https://www.cbsnews.com/news/what-is-deepseek-ai-china-stock-nvidia-nvda-asml/
  51. VentureBeat
    https://venturebeat.com/ai/deepseek-v3-now-runs-at-20-tokens-per-second-on-mac-studio-and-thats-a-nightmare-for-openai/
  52. Hugging Face
    https://huggingface.co/deepseek-ai/DeepSeek-V3-0324
  53. huggingface.co
    https://huggingface.co/deepseek-ai/DeepSeek-R1-0528
  54. China Media Project
    https://chinamediaproject.org/2025/06/12/chinas-global-ai-firewall/
  55. huggingface.co
    https://huggingface.co/deepseek-ai/DeepSeek-V3.1
  56. api-docs.deepseek.com
    https://api-docs.deepseek.com/news/news250821
  57. huggingface.co
    https://huggingface.co/deepseek-ai/DeepSeek-V3.1-Terminus
  58. Native Sparse Attention: Hardware-Aligned and Natively Trainable Sparse Attention
    https://arxiv.org/abs/2502.11089
  59. huggingface.co
    https://huggingface.co/deepseek-ai/DeepSeek-V3.2-Exp
  60. www.cls.cn
    https://www.cls.cn/detail/1672635
  61. ChinaTalk
    https://www.chinatalk.media/p/deepseek-ceo-interview-with-chinas
  62. The New York Times
    https://www.nytimes.com/2025/04/16/technology/nvidia-deepseek-china-ai-trump.html
  63. Machine Decision is Not Final: China and the History and Future of Artificial Intelligence
  64. Rai, Saritha, Loni Prinsloo, and Helen Nyambura "China's DeepSeek Is Beating Out OpenAI and Google in Africa" Bloomberg
    https://www.bloomberg.com/news/features/2025-10-22/china-s-deepseek-pushes-into-africa-making-ai-accessible-to-millions?embedded-checkout=true
  65. High-Flyer
    https://www.high-flyer.cn/blog/3fs/
  66. deepseek-ai/3FS
    https://github.com/deepseek-ai/3FS
  67. High-Flyer
    https://github.com/HFAiLab/hai-platform
  68. DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning
    https://arxiv.org/abs/2501.12948
  69. GitHub
    https://github.com/deepseek-ai/DeepSeek-Coder/blob/main/LICENSE-MODEL
  70. DeepSeek-Coder: When the Large Language Model Meets Programming – The Rise of Code Intelligence
    https://arxiv.org/abs/2401.14196
  71. deepseekcoder.github.io
    https://deepseekcoder.github.io/
  72. deepseek-ai/DeepSeek-Coder
    https://github.com/deepseek-ai/deepseek-coder/
  73. Hugging Face
    https://huggingface.co/deepseek-ai/deepseek-coder-5.7bmqa-base
  74. deepseek-ai/DeepSeek-LLM
    https://github.com/deepseek-ai/DeepSeek-LLM
  75. Math-Shepherd: Verify and Reinforce LLMs Step-by-step without Human Annotations
    https://arxiv.org/abs/2312.08935
  76. DeepSeek-V2: A Strong, Economical, and Efficient Mixture-of-Experts Language Model
    https://arxiv.org/abs/2405.04434
  77. YaRN: Efficient Context Window Extension of Large Language Models
    https://arxiv.org/abs/2309.00071
  78. Hugging Face
    https://huggingface.co/deepseek-ai/DeepSeek-V2-Lite/blob/main/config.json
  79. Hugging Face
    https://huggingface.co/deepseek-ai/DeepSeek-V2/blob/main/config.json
  80. South China Morning Post
    https://www.scmp.com/tech/big-tech/article/3303798/deepseeks-upgraded-foundational-model-excels-coding-and-maths
  81. Hugging Face
    https://huggingface.co/deepseek-ai/DeepSeek-V3/blob/main/config.json
  82. SemiAnalysis
    https://semianalysis.com/2025/01/31/deepseek-debates/
  83. TechSpot
    https://www.techspot.com/news/106612-deepseek-ai-costs-far-exceed-55-million-claim.html
  84. Yahoo News
    https://www.yahoo.com/news/research-exposes-deepseek-ai-training-165025904.html
  85. TheRecursive.com
    https://therecursive.com/martin-vechev-of-insait-deepseek-6m-cost-of-training-is-misleading/
  86. South China Morning Post
    https://www.scmp.com/tech/tech-trends/article/3292507/chinese-start-deepseek-launches-ai-model-outperforms-meta-openai-products
  87. VentureBeat
    https://venturebeat.com/ai/deepseek-v3-ultra-large-open-source-ai-outperforms-llama-and-qwen-on-launch/
  88. TechCrunch
    https://techcrunch.com/2024/12/26/deepseeks-new-ai-model-appears-to-be-one-of-the-best-open-challengers-yet/
  89. Ars Technica
    https://arstechnica.com/ai/2025/01/china-is-catching-up-with-americas-best-reasoning-ai-models/
  90. VentureBeat
    https://venturebeat.com/ai/deepseeks-first-reasoning-model-r1-lite-preview-turns-heads-beating-openai-o1-performance/
  91. The Wall Street Journal
    https://www.wsj.com/tech/ai/china-ai-advances-us-chips-7838fd20
  92. GitHub
    https://github.com/deepseek-ai/DeepSeek-R1/commit/23807ced51627276434655dd9f27725354818974
  93. Reuters
    https://www.reuters.com/technology/artificial-intelligence/deepseek-rushes-launch-new-ai-model-china-goes-all-2025-02-25/
  94. Bloomberg
    https://www.bloomberg.com/news/articles/2025-05-29/deepseek-says-upgraded-model-reasons-better-hallucinates-less
  95. Reuters
    https://www.reuters.com/world/china/deepseek-r2-launch-stalled-ceo-balks-progress-information-reports-2025-06-26/
  96. Financial Times
    https://www.ft.com/content/eb984646-6320-4bfe-a78d-a1da2274b092
  97. Reuters
    https://www.reuters.com/world/china/china-cautions-tech-firms-over-nvidia-h20-ai-chip-purchases-sources-say-2025-08-12/
  98. Nature
    https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12443585
  99. The New York Times
    https://www.nytimes.com/2025/01/28/technology/why-deepseek-could-change-what-silicon-valley-believes-about-ai.html
  100. UC Institute on Global Conflict and Cooperation (IGCC)
    https://ucigcc.org/interview/beyond-the-headlines-on-deepseeks-sputnik-moment-a-conversation-with-jimmy-goodrich/
  101. LCFI - Leverhulme Centre for the Future of Intelligence
    https://www.lcfi.ac.uk/news-events/blog/post/is-sputnik-moment-an-appropriate-analogy-for-the-launch-of-deepseek
  102. Forbes
    https://www.forbes.com/sites/maryroeloffs/2025/01/27/what-is-deepseek-new-chinese-ai-startup-rivals-openai-and-claims-its-far-cheaper/
  103. arXiv
    https://arxiv.org/abs/2412.19437
  104. TIME
    https://time.com/7211646/is-deepseek-panic-overblown/
Image
Source: —
Tip: Wheel or +/− to zoom, drag to pan, Esc to close.