DeepSeek V4 goes official in mid-July with peak-hour pricing that doubles API costs

DeepSeek announced Monday that the production version of DeepSeek V4 will ship in mid-July, bringing performance improvements and a pricing model that charges developers more during peak hours.

The peak/valley system doubles API costs during two daily windows — 9:00 AM to 12:00 PM and 2:00 PM to 6:00 PM — while keeping rates at current levels during off-peak hours. It is a supply-and-demand mechanism designed to smooth traffic spikes and improve service stability, not a blanket price hike.

DeepSeek V4 first appeared as a preview release on April 24, with model weights and code made available under an open license. The model supports a million-token context window and leads Chinese and open-source benchmarks in agent capabilities and world knowledge, with strong results on reasoning tasks. It ships in two variants: V4-Flash for speed-sensitive applications and V4-Pro for compute-heavy workloads.

The timing is notable. Chinese AI labs have been engaged in a prolonged price war, and DeepSeek’s move to introduce time-of-day billing is unusual in the API market — most providers charge a flat per-token rate regardless of when the call is made. Whether developers will accept paying 2x for daytime inference is an open question, but the model already has a substantial user base from its preview phase.