Qianwen Office launches GLM-5 for the first time

Qianwen Office launches GLM-5.3 and DeepSeek V4Pro for the first time, integrating three domestic flagship models. Alibaba's agent product "Qianwen Office" announced on the evening of August 14 that two cutting-edge models, GLM-5.3 and DeepSeek V4Pro, are launched for the first time. From now on, users can directly select them from the "Frontier Model" section on the product homepage. Coupled with the previously connected Qwen3.8-Max, Qianwen Office has become the first office agent product in the industry to cover the three domestic flagship models of Alibaba, Zhipu, and DeepSeek. ZXQKEEP3 ZXQ V4Pro and GLM-5.3 were released on the evening of August 13th and August 14th respectively. They are the latest flagship models of the two companies. DeepSeek V4Pro has 1 million Token contexts and up to 384K output, focusing on strengthening the ability to connect long-range tasks and Agent tasks; GLM-5.3 achieves about 50% performance improvement through post-training based on the base model, and performs well in scenarios such as programming, long-range complex engineering tasks, and network security defense. "Qianwen Office" launched its public beta on August 3. It is integrated with three products: QoderWork, MuleRun, and Wukong. It is the first product in the industry to support desktop Agent, cloud Agent, and enterprise collaboration Agent at the same time. Currently, individual and corporate users can experience it through the official website. The product is also planned to be launched as an independent App and to launch an international version. This access means that Qianwen Office has officially moved towards a multi-model aggregation strategy. Following similar products such as Tencent WorkBuddy and Byte TRAE Work, it further enriches the availability options of domestic models in enterprise-level agent scenarios. via AI News (author: AI Base)