DeepSeek V4 Pro Released: Flagship AI Model Out of Preview

DeepSeek has released the production version of its flagship model, DeepSeek V4 Pro. This ends a preview period that lasted nearly four months. The build, designated V4 Pro 0813, is now the general-availability release. DeepSeek's API documentation lists this version as powering the deepseek-v4-pro endpoint. The API pricing remains consistent with the preview: $0.435 per million input tokens on a cache miss, and $0.87 per million output tokens. It features a one-million-token context window. DeepSeek V4 Pro is a mixture-of-experts system with 1.6 trillion total parameters. The company states its architecture cuts single-token inference compute and KV cache significantly compared to its previous generation. Both V4 models were pre-trained on over 32 trillion tokens. This development matters because it offers a new, powerful AI model now widely available for use.