DeepSeek updated its R1 reasoning AI, and released a more exciting, smaller R1 version that can run on just one GPU.

The tightening of U.S. chip export controls on China has forced Chinese artificial intelligence developers such as DeepSeek to get creative, finding more efficient ways to train AI models on limited computing resources.

A developer reported that Deepseek’s latest AI model has heightened censorship, particularly regarding criticisms of the Chinese government.

DeepSeek has rolled out R1-0528, a major upgrade to the Chinese start-up’s R1 reasoning model, which was released in January.

The Chinese AI model that shook up the industry as a more cost-efficient alternative to the ones from OpenAI, Google, and Meta, now has a new update dubbed DeepSeek-R1-0528. DeepSeek says its latest model has a reduced “hallucination” rate,

For investors, DeepSeek's emergence is causing a serious rethink regarding sky-high valuations of U.S. tech firms, especially those invested in AI.

DeepSeek has gone viral. Chinese AI lab DeepSeek broke into the mainstream consciousness this week after its chatbot app rose to the top of the Apple

DeepSeek triggered a wave of market enthusiasm for Chinese artificial-intelligence when it shocked the market with its AI model in January. Here's what has happened since then.