DeepSeek V4 Pro (Variant 1)
DeepSeek / 1M
DeepSeek·1M context·Text
Flagship MoE LLM with 1.6T total / 49B active parameters and native 1M context for advanced math, logical inference, and specialized coding.