DeepSeek V4 Pro

DeepSeek V4 Pro - 1.6T Sparse MoE with Hybrid Long-Context Attention

DeepSeek V4 Pro is DeepSeek's flagship model in the V4 series: a 1.6 trillion total parameter, 49B active parameter Mixture-of-Experts language model built around a newly designed hybrid attention architecture engineered specifically to make 1 million-token context windows computationally viable.

Architecture

The V4 series introduces three documented architectural innovations over prior DeepSeek generations:

• Hybrid Attention (CSA + HCA): Combines Compressed Sparse Attention and Heavily Compressed Attention across layers to reduce long-context inference cost without degrading comprehension
• Manifold-Constrained Hyper-Connections (mHC): Strengthens residual connections to stabilize signal propagation across the model's deep layer stack while preserving expressivity
• Mixed Precision Training: MoE expert parameters use FP4 precision; most other parameters use FP8

Reasoning Workflow

Three configurable reasoning effort modes per request:

• Non-think - fast responses, no chain-of-thought
• Think High - logical analysis with reasoning
• Think Max (V4-Pro-Max) - maximum reasoning effort; recommended minimum 384K context

Reasoning is surfaced via <think> tags in output.