Oriented on making people closer

Chinese artificial intelligence startup DeepSeek has released an updated version of its reasoning model, R1, significantly intensifying competition with American market leaders such as OpenAI.

DeepSeek introduced the R1-0528 model on the developer platform Hugging Face, although no official public announcement or detailed description of the model has yet followed. Nevertheless, its performance has already attracted the attention of experts.

According to the LiveCodeBench ranking, developed by researchers from the University of California, Berkeley, the Massachusetts Institute of Technology, and Cornell, the updated R1 model from DeepSeek demonstrated impressive results. It trails only slightly behind OpenAI's o4 mini and o3 models in code generation, while surpassing xAI's Grok 3 mini and Alibaba's Qwen 3.

The January launch of the original R1 model had already exerted a significant impact on global markets, triggering a sharp decline in the shares of technology companies outside of China. This precedent also challenged the widely held belief that colossal computing power and investment are required to scale AI.

ORIENT

Photo: ixbt.com