
US bans DeepSeek and AI tools built on its models from federal agencies
The US has banned DeepSeek and third-party AI software built on its models from federal civilian agencies over security concerns.

The US has banned DeepSeek and third-party AI software built on its models from federal civilian agencies over security concerns.

DeepSeek is raising V4 API prices sharply from 16 August, although its latest models remain cheaper than several major rivals.

US escalates accusations that Chinese AI firms used American models without permission, raising tensions ahead of leaders’ talks.

DeepSeek launches new V4 AI models with long context memory and enhanced reasoning amid rising global scrutiny.

Google-backed research suggests advanced AI models reason through internal debate, reshaping how intelligence in machines is understood.

DeepSeek launches Math-V2, the first open AI model to achieve gold-level scores at the International Mathematical Olympiad.

Huawei co-develops DeepSeek-R1-Safe AI model with Zhejiang University, boosting censorship tools and outperforming rivals in safety tests.

DeepSeek’s dominance in China’s AI market is slipping as Alibaba’s Qwen models and other rivals rapidly gain ground, new data shows.

DeepSeek's chatbot experiences a significant decline in downloads as users in China shift to AI apps focused on work, study, and productivity.