DeepSeek, a Hangzhou-based AI lab, said it will raise prices for its V4 models on August 16.

The company will also introduce peak and off-peak rates for output tokens, a measure of how much text an AI model processes or generates.

DeepSeek-V4-Flash will cost US$1.32 per 1 million output tokens during peak hours and US$0.66 off-peak, up from US$0.28.

V4-Pro will cost US$3.96 per 1 million output tokens during peak hours and US$1.98 off-peak, up from US$0.87.

The company said the changes are meant to use computing resources more efficiently and encourage developers and enterprises to shift work to less busy periods.

Even after the increase, DeepSeek’s listed prices remain below the prices it cited for comparison: Moonshot’s Kimi K3 costs US$15 per 1 million output tokens, while Anthropic PBC’s Fable 5 costs US$50.

🔗 Source: Bloomberg

Recent DeepSeek developments