
Alibaba says its new flagship model scored near the top of coding benchmarks and adds safeguards, but is not open weight.
Alibaba has launched Qwen4-Max, saying its new flagship AI model performed near the top of coding benchmarks. The company said it outperformed Anthropic’s Claude Fable 5.1 and Moonshot’s Kimi K3.1 Pro.
Alibaba also said Qwen4-Max surpasses its Qwen3.8-Max model, released in August, while using fewer tokens and offering better cost efficiency. The new model is not open weight, but smaller versions of the series are expected soon.
The company said it added security checks during training and strengthened safeguards against misuse, including in cybercrime and hazardous materials. Alibaba said internal testing found no evidence of significant high-risk capabilities.
This article was produced with the help of AI technology. Based on reporting by the outlets listed above.