Alibaba Qwen3.8-Max reactions: “An API business model wearing an open source jacket”

Alibaba has unveiled Qwen3.8-Max, the most powerful model in the Qwen series, boasting 2.4 trillion parameters and a context window of up to 1 million tokens. This multimodal model is built upon the foundation of Qwen 3.5 and supports long-horizon tasks such as processing extensive codebases or documentation. Alibaba has taken a unique approach by open-sourcing the weights of Qwen3.8-Max, marking the first time they will release this level of openness for a Qwen-Max-class model.

The announcement has garnered mixed reactions, with focus on the number of parameters and the model's performance in various tasks. However, industry experts emphasize that these factors alone do not guarantee a model's competency. Jeff Brokaw, an independent software engineering and AI consultant, warns that Alibaba's release notes only promise the open weights to follow, rather than delivering them immediately.

He cautions that Alibaba's rankings are based on their own evaluation, which may not be a reliable indication of the model's true capabilities. Phil Whittaker, a staff engineer at Umbraco, agrees that benchmarks like those showcased with Qwen3.8-Max will become less important as models become more commoditized. The focus should shift towards quality of the harness, integration with the model, and competing products.

Ajit Dhiwal, a senior staff software engineer at John Deere, questions whether AI companies like Alibaba possess a moat in terms of competitive advantage. He argues that the ease of switching LLMs on the fly, given the abundance of open-source frameworks, diminishes any long-term advantage. Overall, while Qwen3.8-Max represents a significant milestone in the field of AI, the industry remains skeptical of the model's true potential until it undergoes rigorous evaluation by independent parties.