DeepSeek AI
To explore the essence of AGI and benefit all humanity through highly efficient and open-weights large language models.
At a Glance
- AI Developers and Researchers
- Enterprise Software Teams
- Data Scientists
- Global AI Market
AI Tools by DeepSeek AI
(2)
DeepSeek Harness
Open Source Plugin Based Agent Harness
Awesome DeepSeek Integration
DeepSeek Integrations Directory
Discussions
No discussions yet
Be the first to start a discussion about DeepSeek AI
Latest News
DeepSeek Launches 'Harness' Open-Source Coding Agent and V4-Pro Model
DeepSeek Closes Historic $7.4 Billion Funding Round Led by Tencent
DeepSeek-R1 Reasoning Model Released, Sparking Global AI Cost Debate
DeepSeek-V3 Model Released with Competitive Benchmarks
Products & Services
A 671B parameter Mixture-of-Experts (MoE) large language model offering state-of-the-art performance.
A reasoning model designed for deep thinking, matching OpenAI's o1-series in reasoning capabilities at a fraction of the cost.
Latest flagship reasoning model featuring 1.6T total parameters and 49B active parameters, optimized for agentic coding and complex reasoning.
An open-source agentic coding framework with a plugin-based architecture, designed to rival Claude Code and other AI coding assistants.
Market Position
DeepSeek positions itself as a 'Sputnik moment' for AI, providing high-capability reasoning models (R1, V4 Pro) at costs that are 10-34x lower than western competitors like OpenAI and Anthropic.
Leadership
Founders
Liang Wenfeng
Founder and CEO of High-Flyer Capital, a leading quantitative hedge fund in China. Previously worked as an engineer and has a background in AI-driven stock trading since 2008.
Executive Team
Liang Wenfeng
Founder & CEO
Visionary leader and founder of High-Flyer Capital. Pivoted from quantitative trading to AGI research.
Ruan Chong
Chief Scientist
Leading researcher in Mixture-of-Experts (MoE) and multimodal models. Formerly a core contributor to the Janus-Pro project.
Board of Directors
Founding Story
Founded by Liang Wenfeng, a hedge fund entrepreneur who saw the opportunity to pivot from quantitative stock trading to AGI research. DeepSeek was born as an independent project of High-Flyer Capital, utilizing the fund's massive stockpile of GPU computing power to develop frontier AI models at significantly lower costs than US-based competitors.
Business Model
Revenue Model
API usage-based billing (tokens), enterprise services, and likely subscription tiers for web/app products.
Pricing Tiers
Cost-efficient tier for fast responses. $0.14 per 1M input tokens, $0.28 per 1M output tokens.
Flagship reasoning tier with 75% discount currently active. Standard rate is $1.74 / $3.48.
Automatic discount for cache hits on the V4 Pro model.
Target Markets
- AI Developers and Researchers
- Enterprise Software Teams
- Data Scientists
- Global AI Market
- Technical and Coding Tasks
- Complex Mathematical Reasoning
- Enterprise AI Integration
- Document Analysis and Summarization
- Scientific Research Support
- Enterprise users via API
- Global developer community