DeepSeek, a Chinese artificial intelligence startup, said Saturday it cut prices by 75% on its flagship V4-Pro AI model [1].
The move signals an intensifying price war in the global AI market. By lowering the barrier to entry for high-end model usage, the company aims to challenge the pricing strategies of U.S. AI providers [4, 5].
Under the new pricing structure, the cost for the V4-Pro model is now $0.87 per million output tokens [2]. This represents a reduction to a quarter of the original price [2].
Reports on the duration of this discount are contradictory. Some sources said that DeepSeek will make the 75% price cut permanent [1]. However, other reports said the discount was a promotional offer ending on May 5, 2026 [3].
DeepSeek operates out of China, where it has developed the V4-Pro model to compete with Western counterparts [1, 2]. The company's aggressive pricing comes amid a period of heightened tension between the two nations. U.S. officials have frequently accused Chinese AI firms of intellectual-property theft [4, 5].
Industry analysts said that such drastic price reductions are intended to capture market share rapidly. By utilizing specialized hardware, such as Huawei-powered systems, the company may be able to sustain lower operational costs than its competitors [5].
The shift in pricing follows a broader trend of commodity-style competition in the large language model sector. As more providers enter the market, the cost of generating tokens has become a primary battleground for customer acquisition [4].
“DeepSeek announced Saturday a 75% price cut on its flagship V4-Pro AI model”
This pricing strategy reflects a shift from AI as a luxury enterprise service to a commodity. By drastically undercutting U.S. providers, DeepSeek is attempting to force a race to the bottom in pricing, which could pressure Western companies to either lower their margins or innovate more efficient delivery methods to remain competitive.
