Chinese AI startup DeepSeek has quietly released an upgraded version of its R1 reasoning model, intensifying its rivalry with major players like OpenAI and Meta. The enhanced model, posted to Hugging Face, has notable improvements in reasoning, math, and code capabilities and is just behind top OpenAI models in benchmark rankings. DeepSeek previously disrupted the market by outperforming US rivals with a free, open-source AI, leading to sharp declines in US tech stocks and highlighting China’s ability to innovate in AI despite American chip export restrictions. The development comes amid continued discussion about the effectiveness of US technology export controls and China’s rapid progress in artificial intelligence.
Related articles:
OpenAI Launches GPT-5 as Competition Intensifies
China’s Generative AI Startups Defy U.S. Sanctions
