DeepSeek's new model cuts token consumption by over 50%.

DeepSeek has officially launched the new version of its large model, V4.1 Flash-1017. At a time when domestic AI large models are generally in a critical period of iteration, this version has completed its upgrade ahead of others, achieving a significant leap in overall performance. Public data had previously revealed its technical metrics, and this time DeepSeek has for the first time independently released a Pareto frontier curve, visually presenting the model's comprehensive performance across the dimensions of efficiency and cost. V4.1 Flash-1017 reaches an industry-leading level on this curve: its comprehensive cost per unit task is only one percent to one-twentieth that of comparable models such as Fable and Astra, giving it a cost-effectiveness advantage of more than forty to fifty times. Although the released 1017 version maintains the original API pricing system, it has undergone a deep restructuring in Token usage efficiencythrough task-aware compression, dynamic context trimming, and adaptive decoding strategies, it significantly reduces the actual Token consumption of various types of requests. Actual testing shows that in most scenarios Token usage has dropped by more than 50% compared with the previous version, and the corresponding TCO level will likewise see a new round of substantial decline in tandem.