Alibaba touts AI model as better than DeepSeek, OpenAI products, driving stock surge
Alibaba has unveiled a new open-source AI reasoning model, QwQ-32B, that the company claims surpasses the performance of DeepSeek's R1. Despite having significantly fewer parameters (32 billion versus 671 billion), Alibaba's model reportedly matches or outperforms DeepSeek's in key areas like mathematics, coding, and problem-solving. The announcement led to a surge in Alibaba's Hong Kong-listed shares.
The smaller parameter count of Alibaba's model allows it to operate with fewer computing resources, potentially leading to wider adoption. This aligns with Alibaba chairman Joe Tsai's recent emphasis on practical applications in AI development. The release of this model, designed for complex problem-solving, comes shortly after DeepSeek's R1 made waves in the tech industry, highlighting the ongoing competition and rapid advancements in the AI sector.
via SCMP Full Text Feed
Alibaba has unveiled a new open-source AI reasoning model, QwQ-32B, that the company claims surpasses the performance of DeepSeek's R1. Despite having significantly fewer parameters (32 billion versus 671 billion), Alibaba's model reportedly matches or outperforms DeepSeek's in key areas like mathematics, coding, and problem-solving. The announcement led to a surge in Alibaba's Hong Kong-listed shares.
The smaller parameter count of Alibaba's model allows it to operate with fewer computing resources, potentially leading to wider adoption. This aligns with Alibaba chairman Joe Tsai's recent emphasis on practical applications in AI development. The release of this model, designed for complex problem-solving, comes shortly after DeepSeek's R1 made waves in the tech industry, highlighting the ongoing competition and rapid advancements in the AI sector.
via SCMP Full Text Feed