DeepSeek raises API prices sharply and introduces peak/off-peak billing

DeepSeek raised API prices across its V4-Pro and V4-Flash models and replaced its flat per-token rates with a peak/off-peak schedule, effective 16:00 UTC on 16 August 2026. Peak hours were set as 01:00–04:00 and 06:00–10:00 UTC, with off-peak rates priced at half the peak rate. V4-Flash output tokens rose from a flat $0.28 per million to $1.32 at peak and $0.66 off-peak; V4-Pro output rose from $0.87 to $3.96 at peak and $1.98 off-peak. DeepSeek said the change was intended “to allocate resources more reasonably” and to encourage developers to shift workloads to less congested periods.

Increases varied by token category. Output prices rose by roughly 355–371% at peak; cached input tokens, the cheapest category, saw the largest proportional jump, with Caixin Global and other outlets reporting rises of up to about 1,100% on some line items. InfoWorld reported the increase as a response to capacity constraints from demand growing faster than DeepSeek could add compute, rather than a simple margin increase, and noted that even at peak rates the models remained competitive with, and in places cheaper than, equivalent OpenAI and Anthropic offerings.

The move reversed a commitment DeepSeek had made earlier in 2026 to make its discounted introductory rates permanent rather than let them expire as a temporary promotion, and marked a retreat from the aggressive, near cost-price API pricing that had characterised its market entry with V3 and R1. It left DeepSeek still priced well below leading closed-weight rivals, but ended the period in which its API pricing was uniformly the cheapest option among frontier-class models regardless of time of day.