The Chinese lab that made frontier AI cheap and open.
DeepSeek is a Hangzhou-based AI lab founded in 2023 by Liang Wenfeng and funded by the quantitative hedge fund High-Flyer. Its R1 reasoning model shook markets in January 2025 by matching leading Western models at a fraction of the reported training cost. In April 2026 it released DeepSeek-V4 with open weights, including a 1.6-trillion-parameter mixture-of-experts model.
Products and models
—
—
0
— flatDo you support the direction DeepSeek is taking AI?
Do you trust DeepSeek to develop advanced AI responsibly?
The trend line appears after a few days of voting.
Estimated AI ideology
DeepSeek releases its frontier models with open weights under permissive licenses and publishes detailed technical reports. Little is known about its views on safety or governance beyond what its releases show.
AI Revolution's estimate based on publicly available statements, products, and policy positions. Not an objective classification.