DeepSeek
Developers and enterprises seeking cost-effective, high-performance AI models and API access, as well as researchers and enthusiasts interested in open-source AI.
Overview
DeepSeek is a Chinese artificial intelligence research company founded in July 2023 by Liang Wenfeng, who also serves as CEO. The company is headquartered in Hangzhou, Zhejiang, China, and is owned by High-Flyer, a Chinese hedge fund. DeepSeek focuses on exploring the essence of AGI and develops open-source AI models, including the DeepSeek LLM, DeepSeek Coder, DeepSeek Math, DeepSeek VL, DeepSeek V2, V3, R1, and the latest V4 series. The company offers a free conversational chatbot, a web interface, and an API platform for developers. As of 2025, DeepSeek employs around 160 people, with a focus on research over revenue.
Funding: DeepSeek has raised over $7 billion in its first funding round, at a valuation exceeding $50 billion (as of 2026).
What it does
DeepSeek develops and releases open-source large language models and AI systems. Its product lineup includes the DeepSeek chatbot (available as an app and web version) for free conversation, and an API platform that allows developers to integrate DeepSeek's latest models into their applications. The company also publishes research on its models, including technical specifications and training frameworks. Recent models include DeepSeek-V4-Flash, which features enhanced agent capabilities, and DeepSeek-V4-Pro.
Key features
- Open-source AI models (DeepSeek LLM, Coder, Math, VL, V2, V3, R1, V4 series)
- Free conversational chatbot (web and app)
- API platform for model integration
- Agent capability enhancements in V4-Flash
- Transparent pricing per million tokens
- Research publications and technical documentation
Use cases
- Building AI-powered applications via API
- Conversational AI for customer support or personal use
- Code generation and assistance (DeepSeek Coder)
- Mathematical reasoning (DeepSeek Math)
- Vision-language tasks (DeepSeek VL)
- Research and experimentation with open-source models
Pricing
| Plan | Price |
|---|---|
| DeepSeek-V4-Flash | — |
| DeepSeek-V4-Pro | — |
Pricing per 1M tokens. Cache hit input tokens are cheaper (e.g., $0.0028 for V4-Flash). Free chatbot available.
Pricing is gathered from the company's public pages and may change. Check the vendor's site before buying.
Pros and cons
Strengths
- Free access to the chatbot
- Open-source models with published technical details
- Competitive API pricing (e.g., V4-Flash at $0.14 per 1M input tokens)
- Strong backing from High-Flyer and significant funding
- Focus on research and AGI exploration
Limitations
- Pricing not fully transparent on the main website; detailed pricing only in API docs
- No self-serve signup for API; requires developer integration
- Limited language support (primarily Chinese and English)
- Early-stage company with relatively small team (160 employees)
What sets it apart
- Open-source models with frontier performance at lower compute costs
- Owned by a hedge fund (High-Flyer), enabling independent research funding
- Focus on AGI research rather than commercial revenue
Frequently asked questions
Is DeepSeek free to use?
Yes, the DeepSeek chatbot is free to use via the app and web version. API access is paid.
What models does DeepSeek offer?
DeepSeek offers a range of models including DeepSeek LLM, Coder, Math, VL, V2, V3, R1, and the latest V4 series (V4-Flash and V4-Pro).
How much does the DeepSeek API cost?
Pricing is per 1M tokens. For V4-Flash, input is $0.14 and output is $0.28. For V4-Pro, input is $1.74 and output is $3.48. Cache hit input tokens are cheaper.
Who owns DeepSeek?
DeepSeek is owned by High-Flyer, a Chinese hedge fund. It was founded by Liang Wenfeng, who is also CEO.
Where is DeepSeek headquartered?
DeepSeek is headquartered in Hangzhou, Zhejiang, China.
Sources
This profile was compiled from the company's own pages and public web research.
Last researched August 11, 2026.
Manage this listing
Own this company? Update your listing information.