Alibaba’s latest Qwen3.7-Max directly takes the top spot among domestic options. Its programming and agent capabilities look very strong—waiting for real-world testing.

CoinWorld News, according to the latest global large model ranking released by Alibaba Research Institute, Qwen3.7-Max scored 57, ranking first among domestic models, close to the scores of the most powerful models like GPT, Claude, and Gemini. In programming, Qwen3.7-Max can handle tasks ranging from building page prototypes to complex multi-file projects. In office scenarios, the model can integrate various tools and coordinate multi-agent collaboration, automatically completing multi-step workflows, significantly improving real productivity. Additionally, during over 35 hours of fully autonomous experiments, the tool was called over a thousand times, maintaining coherence throughout, and demonstrating cross-framework generalization capabilities across multiple intelligent agent frameworks, capable of seamlessly integrating with various agent frameworks and tools.