Less than a week after the new artificial intelligence platform DeepSeek sent investors spiraling in the United States, another AI model has popped up in China, claiming to be even better.
The new AI model, named Qwen 2.5, comes from the Chinese tech giant Alibaba and is supposedly better than any other model currently on the market, in China and the US.
“Qwen 2.5-Max outperforms... almost across the board GPT-4o, DeepSeek-V3 and Llama-3.1-405B,” Alibaba’s cloud unit said in an announcement posted on its official WeChat account, Reuters reported.
On X, Alibaba’s Qwen account posted stats showing how the model compares to its competitors, claiming to outperform them.
“It achieves competitive performance against the top-tier models, and outcompetes DeepSeek V3 in benchmarks like Arena Hard, LiveBench, LiveCodeBench, GPQA-Diamond,” the post says.
DeepSeek’s AI model, DeepSeek-V3, burst onto the scene three weeks ago and has since made a splash in the world of AI, leaving many to believe that the release of Qwen 2.5-Max was strategic, being that its first day was on the Lunar New Year, when most Chinese people are off work.
The success from foreign competition has also created a stir domestically, as competitors in the world of AI look to upgrade their own models and to stay competitive.
Still, questions remain about the performance capabilities and rise of Chinese developed AI models, as OpenAI accused DeepSeek of using its technology to develop its AI.
“The problem is when someone takes our technology and uses it to build their own product,” a source close to OpenAI told Financial Times on Wednesday.
A Bloomberg report found that security researchers at Microsoft discovered last fall that people with possible links to DeepSeek were harvesting data through OpenAI’s application programming interface.
OpenAI commented on this on Wednesday, noting that they are investigating the potential leak.
“We know that groups in the PRC are actively working to use methods, including what’s known as distillation, to try to replicate advanced US AI models,” an OpenAI spokesperson told The New York Post on Wednesday.
“We are aware of and reviewing indications that DeepSeek may have inappropriately distilled our models, and will share information as we know more.”