DeepSeek officially launched the new big model v4.1 Flash-1017. At a time when large domestic artificial intelligence models are generally in a critical period of iteration, this version was the first to complete the upgrade, and overall performance has improved significantly. Previously, public data showed its technical indicators. This time, DeepSeek also independently released the Pareto Frontier Curve for the first time, to intuitively present the model's comprehensive performance in terms of efficiency and cost. V4.1 Flash-1017 has reached the leading level in the industry in this curve: the comprehensive cost per task is only one-hundredth to one-twentieth of similar models such as Fable and Astra, and the cost performance advantage is more than 450 times. Although the 1017 version released this time maintains the original API price system, it has completed a deep restructuring in terms of token usage efficiency — through task-aware compression, dynamic context cropping, and adaptive decoding strategies, the actual token consumption for various types of requests is significantly reduced. Actual measurements show that in most scenarios, token usage has dropped by more than 50% compared to the previous version, and the corresponding TCO level will also usher in a new round of substantial decline at the same time.

Disclaimer:Webull uses external vendor Google Translation Service for news translations where we endeavour to ensure these are correct, however, we recommend that you please double-check this information accordingly. Webull is not responsible for translation errors or issues.