On April 23, DeepSeek released its V4 model API lineup, including DeepSeek V4-Pro and DeepSeek V4-Flash. Both versions support a native context window of up to 1 million tokens.
V4-Flash is designed for low latency and cost efficiency in real-time applications, while V4-Pro uses a mixture-of-experts (MoE) architecture for tasks such as coding, mathematics, and ...