千问办公首发上线Qwen3.8-Flash，生成速度提升100%，Token消耗减少75%Qwen Work Launches Qwen3.8-Flash, Leading With Generation Speed

Qwen Work (Qianwen Bangong) has launched Qwen3.8-Flash as its first flagship model release, with generation speed positioned as the headline improvement. Flash-tier models trade some ceiling capability for latency and cost, and for an enterprise Agent product the trade is usually worth it: most workplace tasks are throughput-sensitive, not frontier-benchmark-sensitive. By leading with a fast, cheap model inside Qwen Work, Alibaba signals that the product is built for volume usage—copilots that answer quickly inside DingTalk and business flows—rather than occasional hero demos. It's a pragmatic sequencing: make the assistant feel instant at work first, elaborate capability later.