Alibaba Group’s Qwen3.7-Max model secured fourth place on the global Code Arena leaderboard. The model outperformed competing AI products from OpenAI and Google. Alibaba is the only non-US company currently ranked in the top five. Anthropic’s Claude models occupy the other leading positions.
This milestone highlights a strategic shift by Chinese developers toward specialized coding agents. The Code Arena benchmark requires AI to build entire web applications independently from user prompts. Developers vote anonymously on results to rank real-world performance.