DeepSeek splits API pricing into peak and off-peak tiers

DeepSeek introduced time-of-day pricing for its V4 Flash and V4 Pro API models, replacing flat per-token rates with peak and off-peak tiers effective 16:00 UTC. Off-peak rates run fifty percent below peak, with peak hours set at 01:00 to 04:00 and 06:00 to 10:00 UTC, seven hours daily, and off-peak covering the remaining seventeen. DeepSeek announced the change directly; it applies to both the flagship V4 Pro model and the cheaper V4 Flash tier.

Despite the off-peak discount framing, the restructuring amounts to a net price increase: even the cheaper off-peak rate sits above what DeepSeek currently charges across every pricing category, from cached input tokens to output. It follows V4 Pro's move to general availability on August 13, and brings DeepSeek's pricing closer to the variable, demand-based schemes already common among cloud infrastructure providers.

Charging more during peak hours nudges heavy API users toward off-peak scheduling, the same lever cloud providers have long used to smooth infrastructure load rather than simply expanding capacity to meet it.