Forget DeepSeek: Alibaba just unveils new AI model it claims surpasses DeepSeek-V3

Chinese AI startup DeepSeek has sucked the oxygen out of the news cycle over the past week, but it looks like its time at the top might be short-lived. Chinese tech giant Alibaba just rolled out a new version of its Qwen AI model, and it’s making some bold claims—saying its latest release, Qwen 2.5-Max, outperforms DeepSeek-V3, OpenAI’s GPT-4o, and Meta’s Llama-3.1-405B.

Alibaba’s announcement landed on January 29, 2025, coinciding with the first day of the Lunar New Year—an intentional move that signals just how high the stakes have become in the AI arms race.

AI Competition Heats Up

According to Alibaba’s cloud division, Qwen 2.5-Max outshines its competitors across multiple benchmark tests. The market took notice—Alibaba’s U.S.-listed shares jumped 3% after the news, while Microsoft and Meta, both deeply invested in AI, saw slight declines.

This latest development follows DeepSeek’s explosive entrance into the AI scene. The startup’s DeepSeek-V3 and R1 models made waves, pushing bigger players to accelerate their AI rollouts. Now, Alibaba is making sure it’s not left behind, and other major Chinese firms—ByteDance, Baidu, and Tencent—are also scrambling to update their models.

The news of the launch comes just a week after TikTok’s parent company, ByteDance, unveiled Doubao 1.5 Pro and UI-TARS, claiming it outperforms OpenAI’s latest reasoning tools.

Alibaba’s AI Play

Alibaba’s Qwen series is part of a broader effort to establish itself as a major force in AI. The company continues to pour resources into research and development, aiming to make its AI models competitive across different applications.

But the bigger question is—how does Qwen 2.5-Max really stack up? Alibaba insists it has the edge, saying it outperforms DeepSeek-V3 and Meta’s Llama in key tests.

As the AI race intensifies, one thing is clear: No model stays on top for long.

For a deeper look, check out this video: