Chinese Alibaba unveiled a new version of its artificial intelligence model, Qwen 2.5, on Wednesday, claiming that it surpasses the highly regarded chatbot DeepSeek-V3.
Source. This was reported by Reuters.
The timing of the Qwen 2.5-Max release is notable as it coincides with the first day of the Lunar New Year celebrations, a time when most Chinese people are not working.
This highlights the significant impact that the rapid rise of the Chinese AI startup DeepSeek has had on its competitors both in China and abroad.
Alibaba's cloud division announced on its official WeChat account that Qwen 2.5-Max outperforms nearly all aspects of GPT-4o, DeepSeek-V3, and Llama-3.1-405B – the leading models from OpenAI and Meta with open-source code.
The global technology sector is on the brink of revolutionary changes following the introduction of a new AI bot based on the DeepSeek-V3 model by the Chinese startup DeepSeek on January 10.
The unique feature of this bot is its ability to compete with American developments at significantly lower costs, and in some areas, to even exceed them.
Two days after the release of DeepSeek-R1, TikTok's parent company, ByteDance, introduced an update to its flagship AI model, claiming that it surpasses OpenAI's o1, developed with support from Microsoft, based on AIME testing – a primary method for evaluating AI that determines how well models understand complex instructions and respond to them.
DeepSeek also asserts that their R1 model holds its ground against OpenAI's o1 across several benchmarks.
The previous model from the startup, DeepSeek-V2, sparked a price war among AI models in China following its market launch in May 2024.
The fact that the DeepSeek-V2 chatbot has open-source code and is unprecedentedly cheap – 1 yuan ($0.14) for 1 million tokens – led Alibaba's cloud division to report a 97% price reduction across its range of models.
Background. As reported, Nvidia experienced a record drop in market capitalization in a single day amid the success of Chinese DeepSeek. Investors believe that future AI models can be developed with greater efficiency, which will reduce demand for powerful Nvidia graphics processors.