Chinese AI Firm DeepSeek Slashes V4-Pro Model Prices by 75%

Chinese artificial intelligence company DeepSeek has announced a significant and permanent price reduction for its flagship V4-Pro AI model, slashing costs by 75%. The company stated that the API pricing for the V4-Pro model will now fall between 0.025 yuan and 6 yuan per million tokens, depending on the usage type. This is a substantial decrease from the previous rates, which ranged from 0.1 yuan to 24 yuan per million tokens.

While DeepSeek did not explicitly link the price reduction to the increased availability of Huawei's Ascend 950 chips, which the company utilizes to enhance its V4 models' performance, the timing is notable. The pricing change follows broader shifts within China's semiconductor and AI sectors, influenced by U.S. export restrictions on advanced chip sales to China.

These U.S. controls have limited NVIDIA's ability to sell some of its most advanced processors in the Chinese market, consequently opening avenues for domestic chip manufacturers like Huawei. When DeepSeek initially launched the V4 model last month, it highlighted that the Pro version's higher cost was due to constraints in high-end computing capacity. The company had also suggested that prices could decrease significantly once Huawei's Ascend 950 supernodes became more widely available later in the year.