Fri Jan 31 14:30:00 IST 2025: ## Alibaba Launches Qwen 2.5-Max AI Model Amidst DeepSeek Controversy
**BEIJING** – Alibaba has launched its new AI model, Qwen 2.5-Max, on the first day of the Chinese Lunar New Year, a move seen as a direct response to the burgeoning competition in the Chinese AI market. Alibaba claims Qwen 2.5-Max surpasses rival models, including DeepSeek-V3, GPT-4o, and Llama-3.1-405B. However, the launch comes shrouded in uncertainty due to ongoing controversies surrounding DeepSeek’s own performance claims.
Developed by Aliyun’s Tongyi team, Qwen 2.5-Max is a Mixture of Experts (MoE) model trained on over 20 trillion tokens and is reportedly freely accessible to developers via the Qwen Chat platform.
The credibility of DeepSeek, a leading Chinese AI model, is currently under intense scrutiny. OpenAI, the creator of ChatGPT, suspects DeepSeek may have leveraged its proprietary data, potentially employing model distillation techniques to train its model at significantly lower cost than OpenAI’s. Investigations by OpenAI and Microsoft are underway to determine whether DeepSeek used OpenAI’s API, raising concerns about the ethical sourcing of training data within the rapidly developing Chinese AI landscape.
The controversy surrounding DeepSeek casts a shadow over other Chinese AI advancements, including Alibaba’s Qwen 2.5-Max. Experts warn that a lack of transparency in the training processes of these models could damage the credibility of the entire Chinese AI sector in the global race for AI dominance. Until DeepSeek addresses the allegations, skepticism is likely to persist regarding the performance claims of both DeepSeek and its competitors.
First Piper 