In a bold move, Alibaba has unveiled the next iteration of its AI model, the Qwen 2.5-Max, which the Chinese tech giant claims surpasses DeepSeek-V3, the breakthrough model from the rapidly rising Chinese AI startup, DeepSeek.
The timing of the Qwen 2.5-Max’s launch, coinciding with the first day of the Lunar New Year holiday, highlights the mounting pressure DeepSeek’s impressive rise has placed on both international and domestic competitors. As most Chinese citizens take time off to celebrate with their families, Alibaba’s announcement signals a competitive response to DeepSeek’s momentum.
Follow THE FUTURE on LinkedIn, Facebook, Instagram, X and Telegram
According to Alibaba’s cloud division, Qwen 2.5-Max outperforms the latest AI models, including GPT-4, DeepSeek-V3, and Llama-3.1-405B, which are some of the most advanced open-source systems available.
DeepSeek made waves on January 10 with the debut of its AI assistant based on the DeepSeek-V3 model. Shortly after, on January 20, it launched the R1 model, sparking a major reaction in Silicon Valley. The launch sent tech stocks into a downward spiral, with investors questioning the massive spending plans of top US AI firms, given DeepSeek’s much lower development and operating costs.
Adding to the excitement, just two days after DeepSeek’s R1 debut, ByteDance—owner of TikTok—revealed an update to its core AI model. The company claimed it outperformed OpenAI’s GPT in AIME, a complex AI benchmark that evaluates a model’s ability to understand and execute intricate instructions.
This claim mirrors DeepSeek’s assertions that its R1 model is more efficient than OpenAI’s offerings across multiple performance metrics.