The Alibaba Cloud Qwen AI platform (MaaS) has recently updated its model service matrix, with the flagship large model GLM-5.3 from Zhipu and the official version of DeepSeek-V4-Pro now integrated. The corresponding API services are also open to the public, allowing developers to quickly call model capabilities.

The two models launched this time focus on programming and intelligent agent core application scenarios. GLM-5.3, based on Zhipu's latest technological iteration, significantly enhances programming capabilities, long-term task execution capabilities, and network security capabilities; DeepSeek-V4-Pro focuses on improving intelligent agent interaction capabilities, suitable for code development and multi-step complex task processing needs.

QQ20260821-105730.jpg

Currently, the Qwen AI platform has aggregated a variety of mainstream models, covering five major areas: text generation, code development, voice processing, image generation, and video generation. The platform has integrated Tongyi Qwen Qwen3.8-Max, Qwen3.8 open-source models, Qwen Audio3.0, Qwen-Image3.0Pro, Wan3.0, and also introduced domestic mainstream models such as GLM-5.3, DeepSeek-V4-Pro, and Kimi K3. Developers can flexibly select models according to their business scenarios and integrate AI capabilities into their own applications and business systems through APIs.

In terms of payment usage, the platform has launched Token Plan subscription packages. The DeepSeek-V4-Pro, Qwen3.8-Max, and Qwen-Audio series models have opened subscription channels. After purchasing a package, users can freely switch models within AI development tools such as Qoder and Codex.

This new model launch further improves the full-scenario model supply of the Qwen AI platform, offering developers and enterprise users more diverse AI capability options, and helping intelligent agents and code development applications to be implemented.