Alibaba's Qwen 3.7 Max runs 35 hours autonomously with 1,158 tool calls

Alibaba's Qwen 3.7 Max has sparked viral excitement among overseas developers after a demonstration where it ran autonomously for 35 consecutive hours. During this period, the AI agent executed 1,158 uninterrupted tool calls to perform kernel-level optimizations — low-level system improvements typically requiring deep expertise and manual oversight. The model handled complex, multi-step workflows without any human intervention or signs of performance degradation, a stark contrast to most agents that drift or lose coherence on long tasks.

This achievement positions Alibaba as a serious competitor in the global large language model (LLM) race, especially in the agentic AI space where reliability and sustained autonomy are critical. Developers are particularly impressed by the model's ability to maintain context and precision over a 35-hour stretch, suggesting Qwen 3.7 Max could be a viable tool for long-running automation tasks like code review, infrastructure management, and system optimization. The viral reception underscores growing demand for agents that can handle real-world, multi-step workflows without constant human babysitting.