Alibaba’s Qwen3.7-Plus shows how fast the AI race is shifting

Alibaba’s Qwen team, part of Tongyi Lab, has announced Qwen3.7-Plus, a new multimodal AI model. According to GIGAZINE, the model outperformed Claude Opus 4.6 in some benchmark tests. That matters because it shows again that the frontier of AI is no longer shaped only by American companies. Chinese AI labs are moving quickly toward systems that can process text, images, video, interfaces and complex agentic tasks.

Qwen3.7-Plus supports text, image and video input. This places it in a broader transition in artificial intelligence: models are becoming less like chatbots and more like digital agents. GIGAZINE describes Qwen3.7-Plus as a “multimodal interactive hybrid agent” designed for tasks such as recognizing and operating app interfaces, writing code from image input and answering visual questions using knowledge from the internet.

That is a significant shift. The first wave of generative AI was mostly text-based. It answered questions, summarized documents, drafted emails and translated language. The next wave is more operational. These systems can look at screens, understand user interfaces, compare images, generate software and perform actions inside digital environments. In other words, AI is moving from producing responses to carrying out work.

GIGAZINE highlights a demo in which Qwen3.7-Plus uses Python to detect differences between images and solve a spot-the-difference puzzle. Another demonstration shows “desktop app cloning,” where the model writes code for an existing app based on visual input. These examples are not just technical tricks. They point toward a future in which AI can observe a workflow, understand what is happening and then reproduce or automate parts of it.

For companies, pricing is also relevant. Qwen3.7-Plus is available through Qwen Studio and Alibaba’s AI platform. The reported API price is 40 cents per million input tokens, 8 cents per million cached input tokens and 1.60 dollars per million output tokens. That positions the model not only as technically competitive, but also as commercially attractive for high-volume use.

The real importance of Qwen3.7-Plus is that it shows where AI is heading: from conversation to action. For organizations, the key question will no longer be only which model gives the best answer. It will be which model can reliably perform tasks inside real digital workflows.