China's DeepSeek to make permanent 75% price cut on flagship V4‑Pro AI model

Chinese AI company DeepSeek has permanently slashed prices on its V4-Pro model by 75%, making it the cheapest flagship AI model on the market. The new pricing ranges from 0.025 to 6 yuan per million tokens, down from the previous 0.1 to 24 yuan range, depending on usage type. The price cut aligns with DeepSeek’s earlier statement that V4-Pro costs would drop sharply once Huawei’s Ascend 950 supernodes became widely available in the second half of 2026. The company did not confirm whether the permanent reduction is directly tied to increased supply of these chips, which power the model’s performance. Huawei’s AI chip sales have surged due to U.S. export controls blocking Nvidia’s most advanced semiconductors in China, though separate restrictions on chipmaking equipment have limited Huawei’s ability to scale Ascend production. When DeepSeek launched V4 last month, it cited “constraints in high-end compute capacity” as the reason for the Pro version’s higher cost, up to 12 times more expensive than the Flash version. The price adjustment reflects broader industry shifts as Chinese AI firms navigate global tech restrictions and seek cost-effective solutions. DeepSeek’s move could make advanced AI more accessible while highlighting the growing role of domestic hardware like Huawei’s Ascend chips in China’s AI ecosystem.