DeepSeek cuts prices for its flagship V4-Pro model by 75 percent
DeepSeek, one of China's leading AI startups, has permanently reduced prices for its most advanced V4-Pro model by 75 percent. The cost of using this model is now a quarter of its original price. The company attributed this decision to eased constraints on computing power for high-performance systems. This is reported by Ixbt.com .
According to an official statement from DeepSeek, prices for the V4-Pro API service have been lowered to 0.025–6 yuan (approximately $0.0035 to $0.83) per million tokens, depending on the usage type. Previously, this range was 0.1–24 yuan. For context, a token is a unit of text processed by an AI model.
Analysts believe this sharp price drop may be linked to increased shipments of Huawei Ascend 950 chips. Amid US export restrictions on NVIDIA products, Huawei chips have become the primary alternative in the Chinese market. DeepSeek previously noted that the mass deployment of Huawei Ascend 950 super-nodes would allow for significant price reductions.
When the V4 version was introduced last month, the Pro edition was said to be 12 times more expensive than the standard Flash version due to a shortage of computing resources. The price reduction makes DeepSeek models more affordable for developers and businesses, intensifying competition in the AI market.
