Chinese colossus Alibaba released a new version of the model artificial intelligence Qwen 2.5 claiming to exceed Deepseek-V3.

The unusual time of Qwen 2.5-Max release on the first day of the lunar new year, when most Chinese do not work and are with their families, indicates the pressure of Deepseek’s rise in the last three weeks, not only to competitors abroad, but also in its domestic competition.

“Qwen 2.5 Max exceeds almost all GPT-4Os, Deepseek-V3 and Llama-3.1-405B,” Alibaba said in a statement published in its official WeChat account, referring to OpenAi’s most advanced artificial intelligence models Meta.

The release of Deepseek’s artificial intelligence assistant on January 10th, fueled by model Deepseek-V3as well as the movement of the model R1 On January 20, it has shocked Silicon Valley and has caused technology shares fall, as the low cost of Chinese application makes investors question the huge amounts of the top artificial intelligence companies in the United States.

But the success of Chinese Startup Deepseek has led to a race between its Chinese competitors to upgrade their own AI models.

Two days after the release of Deepseek-R1, and Tiktok’s motherboard, byTedance, released update on its top artificial intelligence model, which claimed to overcome OPENAI O1 supported by Microsoft, a report testing how well how well Artificial intelligence models understand and respond to complex instructions.

This echoed Deepseek’s claim that the R1 model was competing with Openai’s O1 in many performance indicators.