The newly launched DeepSeek V4 is available in two variants: a high-end “Pro” model with 1.6 trillion parameters and a lighter “Flash” version built for efficiency. Both feature a one-million-token context window, marking a significant technical leap for handling longer and more complex prompts.