AI 导读
DeepSeek 已将快速、专家与识图三个模式合并升级，不再保留独立的专家选项，并计划在北京时间 9 月 10 日前后上线原生支持多模态的 V4.1 Flash。V4 Pro 服务将于北京时间 9 月 14 日中午停止，此后请求将转由 V4.1 Flash 处理并按后者费率计费。
正文
DeepSeek has unified its Fast, Expert and Image Recognition modes, removing the standalone Expert option.
The company plans to launch V4.1 Flash with native multimodal support around September 10, Beijing time.
V4 Pro service is scheduled to end at noon Beijing time on September 14. Requests will then be routed to V4.1 Flash and billed at its rates—a backend migration alongside the simpler interface.
来源：@thexpin · x.com