Recently, the Qwen Office Platform officially launched the new Qwen3.8-Flash large model, along with a new standard mode. From now on, all users can directly experience the standard mode equipped with this model. Relying on the enhanced model capabilities, users can complete various office tasks with lower point consumption and faster Token throughput efficiency.
According to the official plan, Qwen Office will later form a dual-model supply system of standard mode and advanced mode. The standard mode can cover 95% of daily office needs, while the remaining 5% of complex tasks will be handled by the advanced mode.
The newly launched Qwen3.8-Flash features a new model architecture, achieving strong performance with a total of hundreds of billions of parameters, and its official test results exceed Claude Opus4.6. At the same time, the Qwen large model team has collaborated with the Qwen Office team to create a dedicated version of Qwen3.8-Flash for office use. It focuses on specialized training and optimization for core Agent scenarios such as multi-step planning, tool calling, and context compression; combined with reasoning optimization and a customized Harness architecture, it further improves system throughput efficiency.
Real office scenario test data shows that the standard mode of Qwen Office has increased single-task generation speed by about 100%, and the average Token consumption has decreased by 75%.
For a long time, the AI application field has faced a common industry challenge: pursuing high performance inevitably leads to high costs and high latency, and to control costs, one often has to compromise on the model's intelligence level. Qwen Office, through deep collaboration between Agent and large models, attempts to break the "impossible triangle" formed by performance, cost, and speed.
As the intelligent density of the model continues to improve, and as the platform and underlying model continue to optimize each other, AI Agents are expected to completely free themselves from the anxiety of Token resources, entering a new development stage characterized by high cost-effectiveness and high efficiency - "a lot of resources, well-fed".
