After DeepSeek, Alibaba releases new AI model: Will Qwen 2.5 rival R1?

12punto FOREIGN NEWS

Qwen 2.5-Max was launched after the official nationwide holiday for the Chinese New Year began. This has led to comments that DeepSeek has caused 'panic' not only in the US but also among technology companies operating in China.

In an announcement on its official WeChat account, Alibaba referenced the most advanced open-source AI models from OpenAI and Meta, stating the following:

''Qwen 2.5-Max performs better than almost all of GPT-4o, DeepSeek-V3, and Llama-3.1-405B.''

The release of DeepSeek's AI assistant V3 on January 10 and the R1 model on January 20 caused a decline in the tech stocks of companies operating in Silicon Valley in the US.

The Chinese company's low development and usage costs have led investors to question the high-budget investment plans of leading AI companies in the US.

According to Reuters, DeepSeek's success has also led its domestic competitors to enter a race to develop their own AI models.

Two days after the release of DeepSeek-R1, TikTok owner ByteDance released an update to its own AI model. The Chinese company claimed that it performed better than Microsoft-backed OpenAI's o1 in AIME, a benchmark test that measures how well AI models understand and respond to complex instructions.

AI WARS IN CHINA

The V2 model that DeepSeek released last May triggered a price war among AI models in China.

The fact that DeepSeek V2 was open-source and unprecedentedly cheap (1 Yuan, 0.14 dollars) led Alibaba to implement discounts of up to 97% on its AI models.

Other Chinese tech companies also joined the discount battle. Famous tech brand Baidu and Tencent, which developed the first AI model in China equivalent to ChatGPT, lowered their prices.

In an interview with the Chinese media outlet Waves in July 2024, DeepSeek founder Liang Wenfeng stated that they 'did not care' about price wars at the time and that their main goal was to build AGI (artificial general intelligence).

IS DEEPSEEK'S DIFFERENCE ITS CORPORATE STRUCTURE?

While large Chinese tech companies like Alibaba have hundreds of thousands of employees, DeepSeek operates with a team consisting mostly of young people and PhD students who graduated from China's leading universities.

In his July interview, Liang stated that he thought large tech companies in China might 'not be' in a structure suitable for the future of the AI sector.

Comparing the high costs and hierarchical structures of these companies with DeepSeek's lean operations and flexible management style, Wenfeng noted the following:

'Large foundation (pioneer) models require constant innovation; the capabilities of tech giants are limited.'