War of domination in the field of AI
Chinese technology company Alibaba presented today a new version of the model artificial intelligence of qwen 2.5, for which he claims to outperform the famous Deepseek-V3.
The unusual release of Qwen 2.5-Max, on the first day of the lunar new year, when most Chinese are on holiday and are with their families, marks the pressure exerted by meteorical rise in the last three weeks of Chinese artificial intelligence company Deepseek , not only to its competitors abroad, but also in domestic competition.
“Qwen 2.5-Max exceeds … almost all GPT-4Os, Deepseek-V3 and Llama-3.1-405B,” says the Alibaba Cloud Unit in a statement posted to its official WeChat account, referring to the More sophisticated models of open source artificial intelligence.
The presentation on January 10th of Deepseek’s artificial intelligence assistant, based on the Deepseek-V3 model, as well as the presentation on January 20th of the R1 model, Shocked Silicon Valley and threw shares of technology companiesas the low cost of growth and use of Chinese startup made investors to challenge plans for huge spending from top US artificial intelligence companies.
However, Deepseek’s success also made its domestic competitors upgrade their own models of artificial intelligence.
Two days after the presentation of Deepseek-R1, the owner of TiktokByTedance, presented an update of its own basic artificial intelligence model, for which it claimed to perform better than OPENAI’s Microsoft O1 -backed in Aime, a test that counts how well artificial intelligence models understand complicated instructions and respond to in them.
This reflects Deepseek’s claim that the R1 model outputs several Openai O1 performance criteria.
