Qwen3.7 Plus is a multimodal large language model developed by Alibaba Cloud, released in May 2026. It serves as a unified agent foundation that integrates vision and language processing, designed to function as a multimodal interactive hybrid agent. The model is capable of perceiving real-world scenes, interpreting screen content, and operating graphical user interfaces (GUIs) alongside traditional command-line interactions, building on the text-based reasoning strengths of the Qwen3.7 series.
The model is optimized for "long-horizon" agentic workflows, including autonomous mobile app navigation, multi-step software engineering, and productivity automation. It can generate executable code from visual design references and perform complex reasoning across visual benchmarks such as BabyVision and MathVision. With a context window of 1,000,000 tokens, Qwen3.7 Plus is equipped to process extensive codebases, long-form documents, and high-resolution visual inputs.
Unlike many previous iterations in the Qwen family that featured open weights, Qwen3.7 Plus was introduced as a proprietary model available through Alibaba's cloud infrastructure. It is positioned as a cost-effective alternative to the Qwen3.7 Max model, offering a balance of multimodal capabilities and competitive token pricing. The model supports various agent scaffolds and integrates with the Model Context Protocol (MCP) to facilitate multi-agent orchestration and office productivity tasks.