GMT EIGHT - Latest - DeepSeek's new model cuts token consumption by over 50%.
DeepSeek's new model cuts token consumption by over 50%.
19/10/2026
A Pareto frontier claim of one-hundredth to one-twentieth per-task cost would mark a structural shift in agent economics.
DeepSeek has reportedly 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 reportedly completed its upgrade, with what is described as a significant leap in overall performance. Public data had previously revealed its technical metrics, and this time DeepSeek has released a Pareto frontier curve, visually presenting the model's comprehensive performance across the dimensions of efficiency and cost. V4.1 Flash-1017 reportedly reaches an industry-leading level on this curve: its comprehensive cost per unit task is roughly one-hundredth to one-twentieth the per-task cost of comparable models such as Fable, per the released Pareto curve. Although the released 1017 version maintains the original API pricing system, it has reportedly undergone a deep restructuring in Token usage efficiency, and reportedly reduces the actual Token consumption of various types of requests. Actual testing reportedly 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 reportedly see a new round of substantial decline in tandem.
India's Finance Minister Nirmala Sitharaman blamed the stock market's poor performance on exit mechanisms being too easy and excess liquidity.
Huatai Securities: Diesel prices may remain elevated and are expected to continue to modestly push up US inflation.
Benchmark 10-year Japanese government bond futures rose 0.2 points in morning trading.
Contact US
+852 - 60190728
[email protected]