Português
EnglishSingle task generation speed has nearly doubled, and the average token usage has been significantly reduced by 75%
QwenWork Product
Today, Qwen Office officially became the first to adopt the newly launched Qwen3.8-Flash model and simultaneously introduced a “Standard Service Mode.”
Effective immediately, all users can experience the high performance of Qwen3.8-Flash through the new Standard Mode.
Thanks to advanced model technology, users can complete various work tasks at lower credit costs and with higher token computation efficiency.
Looking ahead, Qwen Office will offer two model specifications: “Standard” and “Advanced.” It is expected that approximately 95% of everyday office needs can be easily handled in Standard Mode, while only a very small portion—about 5%—of highly complex tasks will require Advanced Mode.
This significant leap in user experience is primarily due to upgrades in the model itself, as well as deep collaborative optimization among intelligent agents.
Qwen3.8-Flash employs an innovative architecture with a total parameter count in the hundreds of billions, and its overall performance has already surpassed Claude Opus 4.6.
In addition, the Qwen large model team and the office application team have jointly created a Qwen3.8-Flash version tailored for office scenarios. This version has undergone specialized training and fine-tuning for real-world situations such as multi-step task planning, tool selection, and context compression. Combined with inference optimization techniques and a customized Harness framework, it further enhances processing throughput.
Verified through testing in real office environments, under Qwen Office’s Standard Mode, the generation speed of a single task has nearly doubled, while token consumption has decreased by an average of 75%.
In practical AI applications, high efficiency often comes with high costs and latency, whereas reducing costs may compromise intelligent performance. Now, through close collaboration between intelligent agents and models, the long-standing “impossible triangle” between performance, cost, and speed is gradually being broken.
As model intelligence density continues to increase, coupled with bidirectional adaptation and optimization between the Qwen Office system and the underlying model, AI agents will gradually move beyond token anxiety and enter a new era of abundant resources and stable, efficient operations.
DomTech is DingTalk's official designated service provider in Macau, specializing in providing DingTalk services to a wide range of customers. If you would like to learn more about DingTalk platform applications, please feel free to consult our online customer service or contact us by phone at +852 95970612 or by email at [email protected]. We have an excellent development and operations team with extensive market service experience, ready to provide you with professional DingTalk solutions and services!