Alibaba's Agent product "Qwen Office" announced the launch of two cutting-edge models, GLM-5.3 and DeepSeek V4Pro, on the evening of August 14th. Starting from today, users can directly select these models on the "Cutting-Edge Models" section of the product's homepage. Adding the previously integrated Qwen3.8-Max, Qwen Office has become the first office Agent product in the industry that simultaneously covers the three domestic flagship models from Alibaba, Zhipu, and DeepSeek.

DeepSeek V4Pro and GLM-5.3 were released on the evening of August 13th and August 14th, respectively, both being the latest flagship models from their respective companies. DeepSeek V4Pro features a context length of up to 1 million tokens and a maximum output of 384K, with a focus on enhancing long-term task and Agent task integration capabilities; GLM-5.3 achieves approximately a 50% performance improvement based on its base model through post-training, showing strong performance in scenarios such as programming, long-term complex engineering tasks, and network security defense.

"Qwen Office" started its public beta test on August 3rd, integrating three products: QoderWork, MuleRun, and Wukong. It is the first office Agent product in the industry that supports desktop Agent, cloud Agent, and enterprise collaboration Agent simultaneously. Currently, both individual and enterprise users can experience it through the official website. The product plans to launch an independent app and an international version in the future.

This integration marks Qwen Office's formal shift toward a multi-model aggregation strategy, further enriching the availability options of domestic models in enterprise-level Agent scenarios, following similar products such as Tencent WorkBuddy and Byte TRAE Work.