ALIBABA QWEN3 AI-THE LAST PRAYMENT OF OPEN AI OF CHINA

Alibaba launched the next generation of its large open source language models, QWEN3, on Tuesday, and experts call it another advance in the Board of Artificial Intelligence of China open source.

In a blog post, the Chinese technological giant said that QWEN3 promises improvements in reasoning, the monitoring of the instruction, the use of tools and multilingual tasks, rivaling with other first level models, such as the De Deepseek R1 in several reference points of the industry.

The LLM series includes eight variations that cover a range of architectures and sizes, offering developers flexibility when they use qwen to build AI applications for edge devices such as mobile phones.

Qwen3 is also Alibaba’s debut in the so -called “models of hybrid reasoning”, which says combines the traditional capabilities of LLM with “advanced and dynamic reasoning.”

According to Alibaba, these models can make a transition without problems between a “mode of thought” for complex tasks, such as coding and a “non -thoughtful mode” for faster responses for general use.

“In particular, the QWEN3-235B-A22B MOE model significantly reduces deployment costs compared to other avant-garde models, reinforcing Alibaba’s commitment to high-performance accessible AI,” said Alibaba.

The new models are already available for free for individual users on platforms such as embrace Face and Github, as well as the Alibaba Cloud web interface. Qwen3 is also being used to feed Alibaba’s assistant, quark.

IA analysts told CNBC that QWEN3 represents a serious challenge for Alibaba’s counterparts in China, as well as industry leaders in the United States.

In a statement to CNBC, Wei Sun, the main artificial intelligence analyst in counterpoint investigation, said the QWEN3 series is a “significant advance, not only for its best performance in the class”, but also for several characteristics that point to the “potential for the application of the models.”

These characteristics include QWEN3 hybrid mode of thought, its multilingual support that covers 119 languages ​​and dialects and its open source availability, Sun added.

Open source software generally refers to software in which the source code is available free on the web for possible modification and redistribution. Earlier this year, the Depseek open source model shook the AI ​​world and quickly became a catalyst for China’s space and the adoption of the open source model.

“The Alibaba launch of the QWEN 3 series emphasizes the strong capabilities of Chinese laboratories to develop highly competitive, innovative and open source models, despite the growing pressure of the most tight US export controls,” said Ray Wang, an analyst based in Washington that focuses on the economic and technology competition of the United States-China.

According to Alibaba, Qwen has already become one of the most widely adopted open -source AI models in the world, attracting more than 300 million downloads worldwide and more than 100,000 models derived from the hugged face.

Wang said that this adoption could continue with QWEN3, adding that its performance claims can make it the best open source model worldwide, although even behind the world’s most avant-garde models such as O3 and O4-MINI of OpenAi.

Chinese competitors like Baidu have also rushed to launch new models of AI after the appearance of Deepseek, including making plans to change towards a more open business model.

Meanwhile, Reuters reported in February that Depseek is accelerating the launch of his successor to his R1, citing anonymous sources.

“In the broadest context of the US-China AI race, the gap between US and Chinese laboratories has probably been reduced until a few months, and some could discuss, even just weeks,” Wang said.

“With the last launch of Qwen 3 and the next launch of Deepseek’s R2, it is unlikely that this gap will expand, and can even continue to be reduced.”