Alibaba Debuts Qwen 2.5-Max AI Model That 'Outperforms' DeepSeek
Alibaba has launched the Qwen 2.5-Max AI model, claiming it surpasses competitors like DeepSeek in various benchmarks. This release comes in response to pressures from evolving technology and competitors, particularly as DeepSeek, developed using lower-cost chips, has rapidly gained popularity. The Qwen 2.5-Max reportedly achieved a top score of 89.4 in the Arena-Hard benchmark, which measures AI models based on their responses to human prompts. Additionally, in the MMLU-Pro benchmark, which tests problem-solving abilities at the college level, Qwen 2.5-Max reportedly outperformed DeepSeek and performed similarly to ChatGPT. The launch coincides with the pressures faced by Chinese tech companies, who are vying to remain competitive in the global market. While some industry figures view the rise of models like DeepSeek as a cost-reduction opportunity, concerns about censorship, particularly regarding sensitive topics in China, have also been raised.
Source 🔗