Tech war explodes: Alibaba drops AI bombshell, claims Qwen 2.5-Max beats OpenAI, DeepSeek 

The artificial intelligence race just got fiercer. Chinese e-commerce giant Alibaba has entered the fray with its latest AI model, Qwen 2.5-Max, setting the stage for an intense showdown against domestic rival DeepSeek and global heavyweights OpenAI and Meta. 

Alibaba’s surprise announcement on the first day of the Lunar New Year signals the urgency to counter DeepSeek’s meteoric rise. Over the past three weeks, DeepSeek’s AI breakthroughs have shaken Silicon Valley, causing a dip in tech stocks. The release of DeepSeek’s AI assistant on January 10, followed by its R1 model on January 20, forced competitors to scramble. 

Alibaba’s cloud division claims Qwen 2.5-Max outperforms OpenAI’s GPT-4o, Meta’s Llama-3.1-405B, and DeepSeek’s V3 model “almost across the board.” This direct challenge underscores the growing competition in China’s AI sector, where companies are racing to assert dominance. 

DeepSeek’s rapid success has not only rattled global AI players but also ignited a fierce battle at home. ByteDance, the owner of TikTok, rushed to update its flagship AI model just two days after DeepSeek-R1’s launch. ByteDance claims its latest update surpasses OpenAI’s o1 model in AIME, a benchmark for AI comprehension and response accuracy. 

Founded just 20 months ago in Alibaba’s home city of Hangzhou, DeepSeek has upended the industry with its low-cost, high-performance AI models. The pressure is now on established players like Alibaba, Tencent, and Baidu to accelerate their AI advancements to stay in the game.

However, cracks are appearing in DeepSeek’s rapid ascent. A recent audit by NewsGuard found DeepSeek’s chatbot lagging behind its Western competitors in accuracy and reliability.

Ranked tenth out of eleven AI models tested, DeepSeek’s chatbot managed only 17% accuracy in delivering news and information. It repeated false claims 30% of the time and gave vague or unhelpful answers 53% of the time, resulting in an 83% fail rate. The report raises questions about DeepSeek’s readiness to compete with established AI leaders like OpenAI’s ChatGPT and Google’s Gemini.