DeepSeek V4 Pro
DeepSeek · April 2026
● activeOpen Weightmixture of expertstextAPI Available
Parameters1.6T
Context Window1M tokens
Description
DeepSeek's largest model with 1.6 trillion parameters and a 1 million token context window (roughly 750,000 words — enough to process several novels at once). Combines the architectural innovations from V3 with the reasoning capabilities of R1, representing the frontier of open-weight AI.
Notable Milestones
- ▸Frontier-class open model for code generation
- ▸Million-token context for entire codebase analysis
Key Innovations
Reasoning
ReasoningStructured step-by-step problem solving, often using chain-of-thought or tree-of-thought approaches.
MoE
MoEArchitecture where only a fraction of the model's parameters are active for each input, allowing massive scale with lower compute.
Open Weight
Open WeightModel weights are publicly released but training data/code may not be. Enables fine-tuning but not full reproduction.
Long Context
Long ContextAbility to process very long inputs (100K+ tokens), enabling analysis of entire codebases or books.
Family Tree
Built On
Lineage
Successors (2)
More from DeepSeek
DeepSeek V12024-01 · 67B
DeepSeek V22024-05 · 236B (21B active)
DeepSeek V32024-12 · 671B (37B active)
DeepSeek R12025-01 · 671B (37B active)
DeepSeek-Coder2023-11 · 1.3B - 33B
DeepSeek-Coder-V22024-06 · 236B (21B active)
DeepSeek V4 Flash Vision Exp2026-08-21 · —
DeepSeek V4.1 Flash2026-09-10 · 680B (38B active)