On April 3, Code Arena—a ranking system under the globally recognized large model blind test platform LMArena, which specializes in evaluating AI programming capabilities—announced its latest rankings. Alibaba’s newest large language model, Qwen 3.6-Plus, achieved a global second-place ranking, outperforming international tech giants such as OpenAI and Google, and earning the distinction of being the highest-ranked Chinese model on the list. Qwen 3.6-Plus excelled particularly in the React-specific ranking, which assesses a model’s autonomous coding capabilities in complex web development scenarios, demanding comprehensive engineering thinking and end-to-end development skills. In multiple authoritative evaluations, this model demonstrated superior performance with fewer parameters, setting a new benchmark for the programming capabilities of Chinese-developed models. With this accomplishment, Alibaba climbed to fourth place in the global AI lab rankings, trailing just behind Anthropic, OpenAI, and Google.