Alibaba Qwen3.7-Plus: New AI Agent for Computer Automation

Alibaba has launched Qwen3.7-Plus, a new multimodal AI model designed for screen and coding automation. This model can read interfaces, choose actions, execute steps, and check results across various applications and cloud tasks. What's interesting is how Qwen3.7-Plus acts on screens. It adds native vision input, screenshot perception, and browser automation to the Qwen lineup. This allows it to combine visual perception with agent capabilities to perform tasks. For example, the Qwen team claims a hybrid agent using Qwen3.7-Plus generated over 10,000 lines of code during an eleven-hour app build. It's also claimed to have recreated the macOS Stocks app by parsing the interface and generating code. The new model is priced at $0.40 per million input tokens and $2.40 per million output tokens. This is significantly lower than figures for its language-only counterpart. Benchmark figures show Qwen3.7-Plus scoring 79.0 on ScreenSpot Pro and 70.3 on Terminal-Bench. The bottom line is that Alibaba is now competing in the computer-use AI space, offering a tool that could change how we automate tasks.