Alibaba Group Holding’s new lightweight model Qwen3.8-27B matched much larger near-frontier rivals while remaining able to run on everyday hardware, according to benchmark firm Artificial Analysis, a result that developers have noted as local AI gains momentum.
The Qwen3.8-27B, a model with 27 billion parameters, performed on par with OpenAI’s GPT-5.6 Luna, which Artificial Analysis described as the most cost-efficient model in the US lab’s latest flagship series. OpenAI, which launched the GPT-5.6 family last month, does not disclose the parameter counts of its models.
Alibaba released Qwen3.8-27B’s model weights last Friday, making the underlying parameters that encode the model’s intelligence publicly available.
Qwen3.8-27B performance on the Agentic Index
On Artificial Analysis’ Agentic Index, which measures models’ performance in AI agent-focused workflows, the small Qwen3.8-27B outperformed the GPT-5.6 series’ mid-tier model Terra and Anthropic’s Claude Opus 4.8, the latter of which was released in May.
