Alibaba's Qwen3.7-Plus automates screen and coding with 1M context window

Alibaba has introduced Qwen3.7-Plus, a proprietary multimodal AI agent now available via API on its Bailian platform. This model is designed to tackle screen and coding automation by combining vision and action capabilities. It can read user interfaces, select appropriate actions, execute multi-step sequences, and verify outcomes — all within a single agent. The standout specification is its 1-million token context window, allowing it to process and act upon extensive screen histories, long code blocks, or entire application workflows without losing track of context. This makes it suitable for complex enterprise tasks like automating cloud deployments, testing workflows, or coding refactors.

The model’s release positions Alibaba as a strong contender in the growing market for autonomous AI agents, competing with offerings from OpenAI, Anthropic, and others. By being accessible through the Bailian API, it invites developers to integrate screen-reading and coding automation directly into their tools and pipelines. Practical use cases include automated UI testing, cloud resource management, script generation, and even debugging. For enterprises, this means reduced manual intervention in repetitive GUI tasks and faster development cycles. The 1M context window is particularly significant — it enables the agent to remember entire interaction histories, making long-running automations more reliable and context-aware.