Chinese startup DeepSeek has launched its new DeepSeek-V4 artificial intelligence model series, which is now available in open source. The release includes two versions, one of which features significantly enhanced agent capabilities.
Key points
- The new series includes the DeepSeek-V4-Pro and a more cost-effective version, DeepSeek-V4-Flash.
- The company announced the release via a statement on the social media platform WeChat.
- DeepSeek previously gained international attention in early 2025 with its R1 model, which demonstrated advanced reasoning capabilities.
Why it matters
The release of a new, high-performance AI model from a major Chinese firm is a key indicator of China's ongoing competitive ambitions in the global artificial intelligence sector.
Comments
No comments yet.