DeepSeek V4-Pro peak/valley pricing takes effect August 17 with up to 1100% peak increase
DeepSeek switched to peak/valley pricing on August 17, with peak hours 9:00-12:00 and 14:00-18:00 Beijing time. V4 Pro cached-input price rose as much as 1100%.
From 00:00 Beijing time on August 17, DeepSeek formally enacted its new V4 API pricing scheme. The new scheme uses peak and off-peak tiered pricing, with peak hours running from 9:00 AM to 12:00 PM and 2:00 PM to 6:00 PM Beijing time on weekdays, and off-peak rates at half the peak level.
For V4-Pro, peak-hour input pricing (cache-miss) rose from 3 yuan to 9 yuan per million tokens, a 200% increase, while output rose from 6 yuan to 27 yuan, a 350% increase. Cached-input price jumped from 0.025 yuan to 0.3 yuan, a 1100% increase. Off-peak prices are 4.5 yuan input, 13.5 yuan output, and 0.15 yuan cached input per million tokens.
This is the first time DeepSeek has implemented an official pricing plan since its August 6 warning about price increases, signaling an end to its long-running "price killer" strategy. The company said the adjustment aims to allocate resources more rationally and ease compute pressure. Even at peak rates, V4-Flash output costs about $1.32 per million tokens, still far below Anthropic Claude Fable 5 at $50 and OpenAI GPT-5.6 Sol at $25.