DeepSeek cuts DeepSeek V4 Pro API prices permanently by 75%

DeepSeek has permanently reduced the price of its flagship artificial intelligence model, DeepSeek V4 Pro, by 75 percent, a move that deepens the price war between Chinese and Western AI companies. The Chinese startup announced on May 22, 2026, that the steep discount, originally scheduled to expire on May 31, would remain in effect indefinitely.

Under the new pricing, one million input tokens without a cache hit cost about $0.435, while one million output tokens cost $0.87. Cache hits reduce the input price further, to roughly $0.003625 per million tokens. These rates are roughly one quarter of the model’s original launch prices. In yuan terms, API pricing now ranges from 0.025 yuan to 6 yuan per million tokens, down from 0.1 yuan to 24 yuan.

The price cut makes DeepSeek V4 Pro significantly cheaper than comparable frontier models from Western rivals. By comparison, OpenAI’s GPT 5.5 charges approximately $5 per million input tokens and $30 per million output tokens, while Anthropic’s Opus 4.7 costs around $5 and $25, respectively. Analysts estimate that DeepSeek V4 Pro is now between 20 and 35 times cheaper than premium Western models for certain workloads.