On the evening of April 26, DeepSeek made a groundbreaking announcement: the pricing for input cache hits across its full suite of API services has been slashed to a mere one-tenth of their original launch prices. Additionally, a time-limited discount of 75% is being offered for V4-Pro until May 5. Following this significant price adjustment, the cost per million tokens input (when a cache hit occurs) now stands at an astonishingly low 0.02 yuan for V4-Flash and 0.025 yuan for V4-Pro, marking a new global benchmark for large-scale model pricing.
