Alibaba's Qwen Launches Qwen3.7-Plus Multimodal Agent Model

Alibaba's Qwen big model has launched Qwen3.7-Plus, an upgraded multimodal version of Qwen3.7, designed as a vision-and-language unified agent base. It keeps text, coding, and tool-use abilities while boosting visual understanding, reasoning, and cross-modal tasks.

The model handles image, video, screen, web, and text inputs to complete complex software and office tasks in GUI, CLI, and tool environments. In Vision Arena, it helped Alibaba rank top 5 globally and first in China.

In text tests, Qwen3.7-Plus performs near Max-level models, excelling in coding agents, reasoning, and multilingual tasks. Multimodal tests show big gains in BabyVision, MathVision, ScreenSpot Pro, etc., with stronger visual reasoning and tool-calling.