Qwen

Qwen is Alibaba's flagship AI model family and the technical backbone of its consumer-facing "千问" (Qwen) app. The model line spans multiple parameter sizes and has been open-sourced by Alibaba Cloud since at least 2024, with the 110-billion-parameter Qwen1.5-110B briefly topping domestic open-source benchmarks that spring . More recent iterations include Qwen 3.7 Max, which sacrificed multimodal capabilities to focus on coding and long-horizon agentic tasks .

In product form, Qwen powers an agent-oriented app that integrates deeply with Alibaba's ecosystem—enabling voice-activated task completion through Alipay, Taobao, Fliggy, and Amap . This sits within a broader corporate push to consolidate Alibaba's scattered AI efforts around a single axis: "one company, one model, one product" .

The model's competitive position is strong but narrowly contested. As of mid-2026, Qwen trades blows with Zhipu's GLM for domestic leadership, with benchmark gaps between them reportedly under one percentage point . The author of a recent piece characterized Qwen as "leading" yet perpetually "just a little bit short" of the breakthrough visibility that GLM captured when it became the most powerful freely available model during Fable 5's restricted period .