DeepSeek Now Prices Tokens Like Electricity: 50% Off-Peak Discount and a Spring Boot Pattern to Profit From It
Three days ago I knew exactly what a DeepSeek call cost me. I had wired DeepSeek V4 Pro 0813 into a Spring Boot app with Spring AI, and the math was simple: $0.435 per million input tokens, $0.87 per million output tokens, and a cache-hit rate so aggressive that long agent sessions stayed embarrassingly cheap ( I wrote up the integration ). Then the pricing update landed, and tokens suddenly have…
DeepSeek has introduced peak and off-peak pricing for its tokens, with off-peak rates being 50% lower than peak. This change took effect on August 16, 2026 at 16:00 UTC. The pricing table shows that off-peak rates are 17 out of the 24 hours in a day, while peak hours are limited to early morning and midday UTC. DeepSeek offers three models: V4 Flash and V4 Pro, with varying pricing based on input and output token rates.
Off-peak pricing is significantly cheaper than peak pricing for all models, except for cache hits, which have a 6x to 12x increase during peak hours. The article suggests that developers can shift their workloads to off-peak hours to save on costs, especially for batch processing tasks and interactive chat. A cost-aware scheduling layer can be implemented to determine the best time for different types of workloads, allowing developers to optimize their costs based on the time of day.
Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
