Alibaba releases Qwen3.7-Max agentic LLM with 35-hour autonomous execution

Alibaba formally launched Qwen3.7-Max at the Alibaba Cloud Summit. The proprietary model is optimized for agentic workflows, autonomous coding, long-horizon tasks, and supports up to 35 hours of continuous execution with over 1M token context. It outperforms previous Chinese models and competes with top US models on coding benchmarks. The release strengthens Alibaba's position in the agent era and expands access via Alibaba Cloud Model Studio.