Chinese AI startup DeepSeek has permanently slashed prices for its flagship V4 Pro model by 75%, a move that could intensify competition in the global AI market and further pressure rivals on pricing.
In a statement released on May 22, DeepSeek said that V4 Pro now costs $0.003625 per million input tokens for cache hits, $0.435 per million input tokens for cache misses, and $0.87 per million output tokens.
A token is a unit of text processed by an AI model.
DeepSeek released the V4 series, including the Pro and lighter Flash variants, in April, positioning the models as the beginning of an era of cost-effective one million context length with strong reasoning, coding, and math performance.
The company did not specify the exact reason behind the permanent reduction. However, the announcement comes amid growing expectations around wider availability of Huawei’s Ascend 950 AI chips, which DeepSeek has previously cited as critical to improving the performance and scalability of its V4 models.
When the company unveiled V4 last month, it said the more advanced Pro version was priced as much as 12 times higher than the lighter Flash variant because of “constraints in high-end compute capacity.”
At the time, DeepSeek had indicated that prices would likely “fall sharply” once Huawei’s Ascend 950 supernodes entered large-scale deployment in the second half of the year.
Huawei’s AI chip business has gained momentum as US export restrictions continue to block NVIDIA from selling its most advanced AI chips in China. At the same time, restrictions on chipmaking equipment have constrained Huawei’s ability to rapidly scale production of its Ascend processors.
DeepSeek’s aggressive pricing strategy is expected to increase pressure on AI companies globally as competition intensifies around inference costs and enterprise adoption of large language models.
ALSO READ: Alteryx Inspire 2026: Three Questions Every Data Leader Should Take to Orlando
