DeepSeek has officially launched V4.1 Flash-1017, a new version of its large model, Sina Finance reported. Empirical testing cited by Sina Finance shows token consumption fell by over 50% in most scenarios compared with the previous version. The new version retains the existing API pricing structure. Sina Finance reported that the model's comprehensive cost per task is one-hundredth to one-twentieth that of comparable models such as Fable and Astra.
