DeepSeek releases DeepSeek-V4 Preview
It features 1M context length, open-sourced models including V4-Pro and V4-Flash, with API access available now.
See the latest news and media coverage for DeepSeek. We track all announcements, press releases, and industry mentions in real time, all in one place.
AI research company developing open-source foundation models
deepseek.comLast updated
In short: DeepSeek released low-cost V4 models, faced U.S. data theft allegations, and prepared for an IPO at a potential $75 billion valuation.
It features 1M context length, open-sourced models including V4-Pro and V4-Flash, with API access available now.
Built on V3.1-Terminus, it debuts DeepSeek Sparse Attention for faster, efficient training and inference on long contexts. Now live on App, Web, and API with prices cut by over 50%.
It builds on V3.1-Terminus and debuts DeepSeek Sparse Attention for longer-context efficiency. API is live on App, Web, and API with prices reduced by over 50%.
The update enhances language consistency, addresses user feedback, and strengthens Code Agent and Search Agent performance.
DeepSeek’s new 552-billion-parameter model uses a Mixture-of-Experts architecture designed to balance capability with efficient computing.
DeepSeek plans to hire Yan Wentao, a partner at venture-capital firm GL Ventures, as its first chie...
It claims its V4.1 Flash outperforms mainstays, yet offers a steep discount to the competition...
The new allegations mark a further escalation in US criticism of China's AI industry practices. Earlier this week, three US government agencies, including the National...
Track DeepSeek and your other target companies to get real-time alerts and weekly summaries delivered straight to your inbox.
Browse news for competitors to DeepSeek and other trending companies.