DeepSeek said it will permanently keep V4-Pro API prices at one-quarter of the original rate. This extends a 75% promotional discount that its pricing page set to expire at the end of May.

The listed rates start at 0.025 yuan (US$0.0037) per 1 million cached input tokens, 3 yuan (US$0.44) per 1 million uncached input tokens, and 6 yuan (US$0.88) per 1 million output tokens.

DeepSeek describes V4-Pro as a 1.6 trillion-parameter mixture-of-experts model for reasoning, coding, and long-horizon agent workflows, with a 1 million-token context window.

The move adds to AI price competition in China. Alibaba Cloud previously cut fees on Qwen models by up to 97%, while Baidu lowered prices on newer Ernie models and offered some lighter versions for free.

🔗 Source: Bloomberg

Recent DeepSeek developments