DeepSeek V4 Pro changes API rates August 16

DeepSeek’s V4 Pro API switches to peak and off-peak billing at 16:00 UTC on August 16, with peak rates of $1.32 per million cache-miss input tokens and $3.96 per million output tokens; off-peak rates are half those amounts, according to the official V4 Pro announcement and pricing update. The change gives schedulable batch workloads a 50% discount while making latency-sensitive requests pay for the busy window. Teams already testing time-of-day inference pricing should log UTC request times before the switch—the model name stays the same, but the bill will not.