[Digital Today reporter Chi-gyu Hwang (황치규)] Chinese AI startup DeepSeek is set to unveil its latest large language model in early March, the Financial Times reported on Feb. 28 local time.
It would be DeepSeek’s first new model in more than a year.
According to the report, DeepSeek, headquartered in Hangzhou, is preparing to launch a multimodal model called V4 with image, video and text generation features. The FT reported this, citing two sources familiar with internal matters.
DeepSeek collaborated with Chinese AI chip companies Huawei and Cambricon to develop V4. It is interpreted as an effort to reduce reliance on U.S.-made chips and strengthen China’s domestic semiconductor ecosystem.
DeepSeek drew attention last year as a dark horse threatening U.S. AI companies such as OpenAI by releasing a series of low-cost, high-performance models.
Chi-gyu Hwang
delight@d-today.co.kr