AI Model Provider Cuts Flash Series Prices by Up to 60%

A domestic large-model company has cut the service prices of its flash series models, with the maximum reductions reaching 60% for input cache hits, 33.33% for input cache misses and 11.11% for output pricing, and it has also introduced differentiated pricing for off-peak and peak periods. Jiemian News reported that the move reflects continued declines in the cost of large-model inference services and may further lower the threshold for commercial AI deployment.