DeepSeek V4: New Peak-Valley API Pricing & Features

Chinese artificial intelligence startup DeepSeek is rolling out its DeepSeek V4 model family with a new peak-valley API pricing system. This system aims to better manage computing demand during busy periods. Here's how it works: token prices will double during designated peak periods, while standard API pricing applies during off-peak hours. DeepSeek will provide advance notifications before these billing changes take effect. This encourages developers to shift non-urgent work to less congested times, improving platform efficiency. What's interesting is DeepSeek V4 also brings significant improvements. It builds on its preview release with stronger reasoning, coding, and AI agent capabilities. The company plans to make 1 million-token context windows standard across its V4 lineup. This greatly expands the amount of information the models can process in a single prompt. This new approach could impact how businesses and developers utilize AI services, optimizing costs and resource allocation.