Alibaba unveils a new AI model that it claims outperforms DeepSeek

Chinese tech giant Alibaba has released a new artificial intelligence model, Qwen 2.5-Max. According to the company, it is superior to DeepSeek V3, Reuters reports.

The unusual timing of the Qwen 2.5-Max release, which fell on the first day of the Chinese New Year, when most Chinese people are celebrating, is indicative of the pressure that DeepSeek's rapid growth over the past three weeks has created. This success has affected not only foreign competitors but also local companies.

“Qwen 2.5-Max outperforms ... almost across the board GPT-4o, DeepSeek-V3 and Llama-3.1-405B,” Alibaba's Cloud Division said in a statement posted on the company's official WeChat account.

The release of the DeepSeek AI assistant based on the DeepSeek-V3 model and the new R1 model stunned Silicon Valley and caused a drop in the shares of tech companies. The claimed low costs of developing and using DeepSeek prompted investors to doubt the high expenses of the leaders of the American AI industry.

DeepSeek's success also sparked competition among its local rivals, who began to improve their AI models. Two days after DeepSeek-R1 was released, TikTok's owner, ByteDance, updated its flagship AI model, claiming that it outperformed OpenAI's o1 in AIME, a test that measures the ability of AI models to understand and respond to complex instructions.