Alibaba's Qwen3.7-Max outperforms some ChatGPT and Gemini versions in coding
Alibaba just dropped a new AI model called Qwen3.7-Max, and it's making waves by outperforming some versions of ChatGPT and Google's Gemini in coding tests.
With a score of 1,541 on the Code Arena leaderboard, it ranks fourth globally—right up there with Anthropic's Claude series.
Basically, this model is proving it's one of the best when it comes to handling tough coding tasks.
Proprietary Qwen3.7-Max on Alibaba Cloud
Qwen3.7-Max isn't just fast: it can work solo for up to 35 hours straight and manage more than 1,000 tool interactions without breaking a sweat.
In one test, it optimized code for an AI chip by running 432 kernel tests and more than 1,100 tool calls.
Unlike earlier open-source versions, this one is proprietary and available through Alibaba Cloud, showing they're serious about leading the autonomous coding game.