Alibaba has previewed a new AI model, Qwen3.8-Max-Preview, which the company says is “second only” to Anthropic’s Claude Fable 5. Alibaba reports the model has 2.4 trillion parameters and describes it as one of “the most powerful models available today.”
A technical snapshot: 2.4 trillion parameters
The previewed Qwen3.8-Max-Preview is reported to contain 2.4 trillion parameters — a common metric for model scale and capacity. Alibaba positions this size within the “frontier” class of models, suggesting parity with leading global models and placing Qwen3.8 directly behind Fable 5 in the company’s public statements.
Availability on Alibaba platforms
Alibaba is making Qwen3.8-Max-Preview available through its Token Plan subscription and on its agent platforms Qoder and QoderWork. The announcement was posted on X in conjunction with the World AI Conference (WAIC) in Shanghai.
What this means for China’s AI ecosystem
The release highlights Chinese development moving into the “multi-trillion-parameter” era, a milestone Alibaba frames as narrowing the gap with international model leaders. Explicitly referencing Claude Fable 5 serves to position Qwen3.8 in the ongoing competition over very large-scale models.
Alibaba has so far limited public details to scale and distribution channels, calling Qwen3.8 one of the most powerful models available. Independent benchmarks and forthcoming technical publications will be needed to assess how Qwen3.8’s real-world performance stacks up against Fable 5 and other state-of-the-art models.
