Alibaba Announces Qwen2.5-Max, Claims it's Better than DeepSeek

Alibaba Announces Qwen2.5-Max, Claims it's Better than DeepSeek

Alibaba has unveiled a new large language model called Qwen2.5 Max, claiming it outperforms key rivals such as DeepSeek’s latest AI, GPT-4o from OpenAI, and Meta’s top-tier models. The announcement arrives shortly after DeepSeek introduced its R1 model that surprised many with its high performance and low operating costs.

Key Points

  • Alibaba says Qwen2.5 Max surpasses DeepSeek on multiple AI benchmarks
  • Qwen2.5 Max uses a Mixture-of-Experts approach to handle large-scale tasks
  • DeepSeek’s recent success appears to have intensified domestic competition
  • Alibaba offers Qwen2.5 Max access via its cloud service and Qwen Chat

Alibaba’s cloud division published test results showing Qwen2.5 Max scoring higher than DeepSeek and other leading models on assessments like Arena-Hard, LiveBench, and LiveCodeBench. In a blog post, the Qwen team stated, “Qwen2.5-Max outperforms DeepSeek V3 in benchmarks such as Arena-Hard, LiveBench, LiveCodeBench, and GPQA-Diamond,” while also showing competitive results against GPT-4o.

The timing of Alibaba’s release, so soon after DeepSeek’s announcement, highlights the intensity of the AI race in China. The fact that Alibaba chose a public holiday (the Chinese new year), to launch Qwen2.5 Max, shows the impact that DeepSeek’s rapid rise is commanding from established Chinese tech giants.

In addition to its performance claims, Alibaba notes Qwen2.5 Max employs techniques like supervised fine-tuning and reinforcement learning from human feedback to improve quality and efficiency. Users can access it via Qwen Chat or via the API. We have also embedded a Hugging Face Demo below:

DeepSeek’s recent splash underscores how lean AI startups can challenge industry mainstays. The Chinese startup reportedly operates with a smaller team yet produces high-performing models at lower costs, putting extra pressure on larger firms. Still, Alibaba remains confident in its capacity to scale Qwen2.5 Max further, signaling a continued emphasis on research and development in China’s fast-moving AI sector.

Chris McKay is the founder and chief editor of Maginative. His thought leadership in AI literacy and strategic AI adoption has been recognized by top academic institutions, media, and global brands.

Let’s stay in touch. Get the latest AI news from Maginative in your inbox.

Subscribe