Alibaba releases Qwen3.7-Plus multimodal agent

Alibaba released Qwen3.7-Plus, a multimodal AI agent combining visual perception with coding and tool use. It recognizes real-world scenes, operates graphical interfaces, writes code from visual templates, and navigates mobile apps end to end. The agent built an English vocabulary app with over 10,000 lines of code in more than 1,000 calls over eleven hours. It recreated macOS Stocks app by parsing UI, generating SwiftUI code, connecting to stock API, and running ten functional tests.