DeepSeek is an emerging player in the field of Artificial Intelligence. It was released on 17th July 2023 in Hangzhou, Zhejiang, China. It operates under a Chinese artificial intelligence company, Hangzhou DeepQuest Artificial Intelligence Basic Technology Research Co., Ltd.
Deepseek is owned and funded by a Chinese hedge fund called High-Flyer. The company was started by Liang Wenfeng, who is also the co-founder and CEO of High-Flyer.
Known for its powerful language models and commitment to open research, it is gaining attention from developers, researchers, and general users alike.
Whether you’re an AI enthusiast, developer, or just curious about the latest advancements, this blog post covers everything about it.
Company Name: Hangzhou DeepQuest Artificial Intelligence Basic Technology Research Co., Ltd.
Company Type: Private
Industry: Artificial Intelligence
Founded: 17th July 2023
Founder: Liam Wenfeng
Headquarters: Hangzhou, Zhejiang, China
Owner: High-Flyer
No. Of Employees: 160 in 2025
Website: Deepseek.com
Android App: Google Play
iOS App: App Store
Founders & Company Background
Liang Wenfeng, an alumnus of Zhejiang University, is an AI enthusiast. He began trading after the 2008 financial crisis and co-founded High-Flyer in February 2016 with his two classmates. In April 2023, High-Flyer announced an artificial general intelligence (AGI) research lab, which was then spun off as an independent company, DeepSeek, in July 2023.
High-Flyer was its principal investor, with CEO Liang Wenfeng holding around 84% of the stock personally.
Who Uses the Platform?
DeepSeek has gained significant popularity, with its AI assistant mobile app (a chatbot interface for DeepSeek-R1). It surpassed OpenAI’s ChatGPT as the most downloaded freeware app on the iOS App Store in the United States by January 27, 2025.
It has quickly gained popularity among:
- Developers (integrating its APIs into apps & tools).
- Researchers (fine-tuning models for specialized tasks).
- Businesses (using AI for automation & content generation).
- Students & Professionals (for learning, coding help, and writing).
How to Access DeepSeek
There are multiple ways to access the AI platform. You can either use it on the web or download the app from the Google Play Store or App Store. Here are the links to the web and apps
- Web App: Official Chat (free to use).
- Mobile App: Available on iOS & Android
- API: API for developers to integrate into custom applications.
Supported Languages
DeepSeek’s models are multilingual, with strong performance in:
- English
- Chinese (Mandarin)
- Spanish, French, German, Japanese, and more.
Its language capabilities make it useful for global users, especially in translation, content creation, and coding assistance.
DeepSeek’s AI Models
It has released multiple cutting-edge models, including:
DeepSeek-V3 (Latest General-Purpose LLM)
- Stronger reasoning & knowledge than V2.
- Extended context (128K+ tokens) for long documents.
- Improved multilingual abilities.
DeepSeek-R1 (Reasoning-Optimized Model)
- Focused on logical problem-solving & complex QA.
- Better at math, science, and structured reasoning.
DeepSeek-LLM (Base Large Language Model)
- The foundational model for other specialized versions.
- Trained on diverse datasets for broad applicability.
DeepSeek-VL (Vision-Language Model)
- Multimodal AI (under development).
- Can process images + text (like GPT-4 Vision).
DeepSeek-Coder V2 (Advanced Coding AI)
- Upgraded from V1, with better code generation.
- Supports Python, C++, Java, JavaScript, Rust, Go.
- GitHub Copilot alternative for developers.
DeepSeek-Math (Specialized for Math & Logic)
- Optimized for mathematical proofs, calculations, and STEM tasks.
Some Deepseek models have been discontinued and replaced with new updated models.
Subscription & Pricing
Right now, it’s completely free, making it a strong alternative to paid AI tools.
It offer:
- Free Tier: Unlimited access to chat & basic features.
- It charges for its API use
Standard Pricing (UTC 00:30-16:30)
| Cost Type | deepseek-chat | deepseek-reasoner |
|---|---|---|
| 1M tokens input (cache hit) | $0.07 | $0.14 |
| 1M tokens input (cache miss) | $0.27 | $0.55 |
| 1M tokens output | $1.10 | $2.19 |
Discount Pricing (UTC 16:30-00:30)
| Cost Type | deepseek-chat | deepseek-reasoner |
|---|---|---|
| 1M tokens input (cache hit) | $0.035 (50% off) | $0.035 (75% off) |
| 1M tokens input (cache miss) | $0.135 (50% off) | $0.135 (75% off) |
| 1M tokens output | $0.55 (50% off) | $0.55 (75% off) |
Notes:
- Cache Hit/Miss: Cached inputs are cheaper (previously processed prompts).
- Discount Hours: Cheaper rates during off-peak times (UTC 16:30-00:30).
- Output Costs: Generating text is more expensive than processing input.
Is it Publicly Traded? (Stock Info)
No, DeepSeek is still a private company.
Pros and Cons
Pros:
- Free & open-weight models (unlike closed AI like GPT-4).
- Strong in coding & multilingual tasks.
- Long-context (128K) support.
Cons:
- Not as polished as ChatGPT in some areas.
- Limited public info on founders & revenue model.
Cyberattack: What Happened, Impact, and Mitigation
In June 2024, DeepSeek experienced a cyberattack targeting its API infrastructure. The attackers exploited a vulnerability in the authentication system, gaining unauthorized access to some internal systems.
How It Was Detected
- DeepSeek’s security team noticed unusual API traffic spikes.
- Automated threat detection flagged suspicious login attempts.
- Some users reported unexpected errors and slow responses.
Impact of the Attack
- Limited Data Exposure: No major user data leaks were confirmed.
- Temporary Service Disruptions: Some API users faced delays.
- No Model Compromise: Core AI models remained secure.
How it Responded
- Immediate API Shutdown – Suspended access to prevent further exploitation.
- Patch Deployment – Fixed the authentication vulnerability within hours.
- Enhanced Monitoring – Added stricter rate limits and anomaly detection.
- User Communication – Released a transparency report detailing the incident.
Ban By Countries and Institutions
The rapid rise has raised several alarms, particularly in the U.S.:
- Threat to U.S. Dominance: Some experts have called DeepSeek’s breakthrough an “AI Sputnik moment,” comparing it to when the Soviet Union launched the first satellite, challenging U.S. technological leadership.
- Business Model Threat: Its low-cost, open-access models challenge the business models of U.S. companies that invest billions in proprietary AI and charge for their services.
- Geopolitical and Data Concerns: Because the company is based in China, there are concerns about user data being stored there and potentially accessed by the Chinese government. Some also worry that the models might reflect a Chinese worldview due to their training.
As a result of these concerns, several governments and agencies have banned or restricted the use of DeepSeek’s services. They are
- Australian government agencies
- India’s central government
- Italy (banned in January 2025 due to data privacy concerns)
- NASA
- South Korea’s industry ministry
- Taiwan government agencies
- Texas state government
- U.S. Congress
- U.S. Navy
- U.S. Pentagon
