
After the company claimed to have developed a new reasoning model which they argued is comparable to global sensation DeepSeek’s R1, the Hong Kong listed shares for Alibaba Group rose by more than 8% on Thursday.
Qwen, the e-commerce titan’s Artificial Intelligence branch, asserted on X that their QwQ-32B model, which has 32 billion parameters, can perform at par with DeepSeek’s R1 model which has 671 billion parameters.
This announcement followed China’s promise to increase funding in the support of more advanced technologies such as AI, humanoid robotics, and 6G telecom a day prior. Government agencies alongside smaller businesses in China are starting to adopt AI models more readily.
According to Alibaba, the new model can be accessed through their chatbot service, Qwen Chat. Qwen Chat users can select multiple models including Qwen2.5-Max, the strongest language model in the Qwen series.
The firm claimed that the QwQ-32B passed benchmark tests for mathematical reasoning, coding, and general problem solving at parent rates to top models like OpenAI’s o1 mini and DeepSeek’s R1.
DeepSeek has become the latest poster child for Chinese AI, taking on top model from OpenAI at a fraction of the costs and less powerful computing.
According to experts, the steps the government has taken will increase adoption of artificial intelligence within the country.
“China is swiftly constructing an application-focused AI ecosystem that goes beyond research; this is about real, economic value,” said Sun Wei, principal AI analyst at Counterpoint.
Alibaba's stock rose 8% to HK$140.5 per share.