DeepSeek permanently lowers V4-Pro AI model prices by 75%
Chinese artificial intelligence startup DeepSeek will make permanent a 75 per cent price reduction. The company announced on Saturday a reduction in the pricing of its flagship V4-Pro artificial intelligence model, maintaining prices at a quarter of their original level. DeepSeek did not reveal whether the permanent price reduction was a result of heightened supply of Huawei’s Ascend 950 chips, which were utilised to optimise V4’s performance.
The company reduced V4-Pro API costs to a range of 0.025 to 6 yuan per million tokens, depending on usage type, down from a previous range of 0.1 to 24 yuan, according to the statement. A “token” is a unit of text that is processed by the AI model. Huawei’s AI chip sales have seen a boost due to U.S. export controls that restrict Nvidia from marketing its most advanced semiconductors in China. However, additional restrictions on chipmaking equipment exports have constrained Huawei’s capacity to enhance Ascend production.
When DeepSeek launched V4 last month, it stated that the Pro version would be priced at up to 12 times more than the less powerful Flash version, citing “constraints in high-end compute capacity” as a factor limiting availability. It was also noted that Pro pricing is anticipated to decline significantly following the large-scale launch of Huawei Ascend 950 supernodes in the latter half of the year.
