Alibaba’s lightweight Qwen model takes on larger AI systems from OpenAI, DeepSeek, Zhipu
Alibaba Group Holding’s new lightweight AI model Qwen3.8-27B has matched much larger near-frontier rivals while being able to run on everyday hardware, impressing developers as local AI gains momentum. The Qwen3.8-27B, a small model with 27 billion parameters, performed on par with OpenAI’s GPT-5.6 Luna, which was billed as the most cost-efficient model in the US lab’s latest flagship series,…
Alibaba Group Holding has introduced a lightweight AI model named Qwen3.8-27B, which has matched larger rivals such as OpenAI’s GPT-5.6 Luna and Chinese firms DeepSeek-V4-Pro-0813 and Zhipu’s GLM-5.2. The smaller model, with 27 billion parameters, was able to compete with more powerful models running on everyday hardware. Artificial Analysis Intelligence Index reported that Qwen3.8-27B outperformed GPT-5.6 mid-tier model Terra and Anthropic’s Claude Opus 4.8.
The model allows developers to run an LLM on a reasonably specced laptop, according to Simon Willison, a British programmer. While the model "defaults to wildly overthinking things" and seems slow in performance, it is a "miracle" that a 17GB file can do everything needed from an LLM for real work, Willison stated. The Qwen3.8-27B release follows a trend of smaller open-weight models becoming increasingly powerful, and global developers' interest in running AI locally on their devices is expanding.
Alibaba's Qwen models have seen over 2 billion downloads globally and are open sourced, with more than 460 models available and 3 billion downloads in total.
Written by urgent.news from SCMP Tech's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
