BEIJING, China: Chinese artificial intelligence startup DeepSeek said on May 23 that it will permanently maintain a steep 75% reduction in prices for its flagship V4-Pro AI model, signaling intensifying competition in China's fast-growing AI sector.

The company said in a statement that prices for the V4-Pro model would remain at roughly one-quarter of their earlier levels following an initial temporary reduction introduced after the model's launch.

DeepSeek did not specify whether the permanent discount was linked to increased availability of Huawei's Ascend 950 AI chips, which the company has used to improve the performance of its V4 series models.

The announcement comes as Chinese AI firms race to lower costs and expand access to advanced models while navigating U.S. export restrictions on high-end semiconductors.

According to the company, API pricing for V4-Pro has been reduced to between 0.025 yuan and 6 yuan per million tokens, about $0.0035 to $0.83 depending on usage type, down from previous pricing of between 0.1 yuan and 24 yuan per million tokens.

A token refers to a unit of text processed by an AI system during tasks such as generating responses or analyzing information.

DeepSeek launched its V4 model series last month and had initially said the more advanced Pro version would cost up to 12 times more than the lighter Flash version because of what it described as "constraints in high-end compute capacity."

At the time, the company also indicated that prices could fall sharply once Huawei's Ascend 950 supernodes became available in larger quantities later this year.

Huawei has emerged as a major beneficiary of U.S. export restrictions that limit Nvidia's ability to sell its most advanced AI chips in China. Chinese technology companies have increasingly turned to Huawei's domestically produced AI processors as alternatives.

However, Huawei's ability to expand production has also been constrained by separate U.S. restrictions on advanced chipmaking equipment exports to China.

DeepSeek's pricing move highlights the growing pressure among AI companies to reduce costs as businesses adopt increasingly sophisticated AI systems.

The company's aggressive pricing strategy could intensify competition with rivals offering large language models and AI infrastructure services in both China and overseas markets.

The announcement also reflects broader efforts by Chinese AI firms to build self-reliant technology ecosystems amid continuing tensions between Washington and Beijing over semiconductors and advanced computing technology.