Alibaba Cloud's Coding Plan unveils four top-level open-source models.

On February 25th, the official account of Alibaba Cloud announced that Alibaba Cloud's Coding Plan subscription service has fully launched four top open-source models: Qwen 3.5, GLM-5, MiniMax M2.5, and Kimi K2.5. Users who subscribe to the package can freely switch and utilize these models on AI tools such as Qwen Code, Claude Code, Cline, and OpenClaw, enjoying more stable models with higher token limits. Currently, Alibaba Cloud is the only cloud provider in the world offering this service. Alibaba Cloud's Coding Plan Lite basic package allows up to 18,000 requests per month, with new users only needing to pay 7.9 RMB for the first month. The Pro advanced package can handle up to 90,000 requests per month, completing thousands of complex tasks, with new users only paying 39.9 RMB for the first month, significantly reducing costs for high-frequency coding demands. The four models launched this time are all top open-source models. Among them, Qwen 3.5, a native multi-modal model open-sourced on New Year's Eve, excels in efficiency and cost-effectiveness, with only 170 billion out of its total 397 billion parameters activated, showing outstanding performance in various benchmark evaluations including inference, programming, and Agent intelligence. GLM-5 is the latest flagship open-source model of KNOWLEDGE ATLAS, with API services being in high demand since its launch; MiniMax M2.5 is a model designed for Agent scenes, performing excellently in Excel advanced processing, deep research, and Office scenarios such as PPT; Kimi K2.5 has also made significant improvements in programming and Agent capabilities. Currently, Alibaba Cloud's Coding Plan has launched eight top programming models, and Alibaba Cloud's BaiLian platform has also launched over 100 mainstream model APIs from domestic and international sources, as well as over 400 AI hardware, short video, advertising content, and other areas of Agent templates and services, meeting the model needs of more enterprise scenarios.