DeepSeek raises API prices up to 1,100% as it launches V4-Pro model
DeepSeek confirmed peak/off-peak API pricing taking effect August 16 with increases up to 1,100%, timed to the launch of its 1.6-trillion-parameter DeepSeek-V4-Pro model.
DeepSeek will raise its application-programming-interface (API) prices by as much as 1,100% under a new peak/off-peak billing structure that takes effect August 16, 2026, according to the Chinese AI lab’s official pricing page. The new rates start at 16:00 UTC, the same day DeepSeek launched DeepSeek-V4-Pro, a 1.6-trillion-parameter model.
Under the new rates, DeepSeek-V4-Pro peak output rises to $3.96 per million tokens, up from a flat $0.87, while off-peak output is set at $1.98. Peak hours run 01:00-04:00 and 06:00-10:00 UTC. The increases reportedly range from about 50% to 1,100% depending on the model, token type, and time of day.
The move inverts the low-price strategy that made DeepSeek a disruptive force in the API market. Time-of-day pricing pushes cost-sensitive workloads into off-peak windows and rewards heavy cache use, a structure closer to how cloud providers meter capacity than the flat token rates most AI vendors advertise.
DeepSeek paired the hike with DeepSeek-V4-Pro, a mixture-of-experts model with a 1-million-token context window, adjustable reasoning effort, and support for OpenAI’s Responses API format. The pricing and the launch finalize a price-increase warning DeepSeek issued around August 6.
Whether developers absorb the change or route around it is the open question. DeepSeek’s appeal has rested heavily on undercutting US rivals; a 1,100% peak premium narrows that gap on exactly the workloads that run during business hours.
An entrepreneur with over a decade of experience in AI, Cloud, and HPC. He is currently a DevOps Architect and the founder of Data Phoenix, an influential media voice for the AI industry, with a strong focus on community building and open source.
More news

AWS makes Amazon Textract Insights generally available in two US Regions

Together AI adds Tencent Hy4 to its serverless API

JAX 0.12.1 moves selected Pallas APIs to `jax.pallas`
