China's Alibaba releases new AI model that allegedly outperforms DeepSeek's V3 and Meta's Llama
Alibaba Group has published test results of the Qwen 2.5 Max AI model and stated that it demonstrates world-leading performance.
Alibaba Group has published test results of the Qwen 2.5 Max AI model and stated that it demonstrates world-leading performance.
In particular, the updated version of Qwen 2.5 Max showed better results in various tests than Llama from Meta Platforms Inc. and the V3 model from another Chinese company, DeepSeek, Bloomberg reports.
Along with Tencent Holdings Ltd. and Baidu Inc., Alibaba has invested significant resources in its cloud services segment and is recruiting Chinese AI developers to use its tools.
As Reuters notes, Alibaba chose an interesting time for its big release — the first day of the Lunar New Year, when most Chinese people are off work and spending time with their families. This indicates that the pressure caused by the rapid growth of Chinese AI startup DeepSeek is affecting not only Western competitors but also domestic competition.
Only on January 10, DeepSeek released an AI assistant based on the DeepSeek-V3 model, and on January 10, it introduced the R1 model, which shocked Silicon Valley with its supposedly low training costs and caused tech stocks to fall. This forced investors to question the huge spending plans of leading AI companies in the US.
DeepSeek’s success has also led to its Chinese competitors rushing to improve their own artificial intelligence models.
Two days after the release of DeepSeek-R1, TikTok owner ByteDance released an update to its flagship AI model, which it claims outperformed Microsoft-backed OpenAI’s o1 in AIME, a benchmark test that measures how well AI models understand and respond to complex instructions.
While large Chinese technology companies like Alibaba have hundreds of thousands of employees, DeepSeek operates as a research lab staffed mostly by young graduates and doctoral students from leading Chinese universities.
In July, DeepSeek’s enigmatic founder, Liang Wenfeng, said he thought China’s biggest tech companies were not well-suited for the future of the AI industry, due to their high costs and vertical management structures. DeepSeek, by contrast, operates on a lean, hands-off approach. «Big fundamental models require constant innovation, and there are limits to what tech giants can do,» he said.
Recall that OpenAI said that Chinese companies are «constantly» trying to use American competitors to improve their AI models. Earlier, the head of OpenAI Sam Altman commented on the hype around the artificial intelligence models from China’s DeepSeek. He called them «amazing», but is confident that his company can do better.
Hype around Chinese AI
DeepSeek, a Chinese artificial intelligence lab funded largely by High-Flyer Capital Management, recently gained popularity after its chatbot rose to the top of the Apple App Store charts .
DeepSeek’s language models, which were trained using efficient computational methods, have led many analysts and technologists in the West to wonder whether the US will be able to maintain its lead in the AI race and whether demand for AI chips will persist.
This excitement has caused stock prices in technology companies to fall, including top graphics processor manufacturer Nvidia, and Mark Zuckerberg hastened to announce that Meta plans to invest $60 billion in AI development by 2025.
The hype around DeepSeek became so great that even monobank co-founder Oleg Gorohovsky and the head of the Ministry of Digital Affairs, Mykhailo Fedorov, wrote cryptic, laconic posts about the startup. The latter later published a more detailed post on Telegram.
dev.ua did a detailed analysis of how DeepSeek managed to outperform its competitors.
DeepSeek has also released a new set of multimodal AI models, which the Chinese company claims can outperform OpenAI’s DALL-E 3 image generator.
ChatGPT is reducing the cost of using the Plus version due to the popularity of the Chinese DeepSeek.
Read the country’s main IT news in our Telegram
Read the country’s main IT news in our Telegram
«The data of millions of Italians is at risk»: Italy became the first country to officially contact DeepSeek regarding the processing of its citizens' data
«The data of millions of Italians is at risk»: Italy became the first country to officially contact DeepSeek regarding the processing of its citizens' data
OpenAI accuses China’s DeepSeek of using US AI models to train chatbot. The company found evidence of model «distillation»
OpenAI accuses China’s DeepSeek of using US AI models to train chatbot. The company found evidence of model «distillation»
ChatGPT lowers the cost of using the Plus version due to the popularity of the Chinese DeepSeek. What will the prices be and who will be able to use it?
ChatGPT lowers the cost of using the Plus version due to the popularity of the Chinese DeepSeek. What will the prices be and who will be able to use it?
