Chinese AI start-up DeepSeek has unveiled a major upgrade to its R1 model, claiming the new version, R1-0528, now rivals top-tier systems from OpenAI and Google.

Announced Thursday, the updated model features significant improvements in reasoning and creative writing, enabling it to produce essays, fiction, and prose with a style reportedly close to that of human authors.

DeepSeek also highlighted stronger coding capabilities, making the model more adaptable for both technical and creative tasks.

The original R1 model launched in January. With this latest update, DeepSeek is positioning itself to better compete with leading Western AI developers, both in China and internationally.