RunLocalAI
deepseek
745B parameters
Commercial OK

DeepSeek V4

DeepSeek's spring 2026 frontier MoE. 745B total / 38B active. The current open-weight benchmark leader on coding + math; closes the gap with closed-source flagships on reasoning.

License: DeepSeek License·Released Mar 15, 2026·Context: 131,072 tokens

Overview

DeepSeek's spring 2026 frontier MoE. 745B total / 38B active. The current open-weight benchmark leader on coding + math; closes the gap with closed-source flagships on reasoning.

Family & lineage

How this model relates to others in its lineage. Family members share architecture and training-data roots; parent / children edges record direct distillation or fine-tune relationships.

Parent / base model
DeepSeek V3 (671B MoE)671B
Frontier
Family siblings (deepseek-v)
DeepSeek V3 Lite (16B MoE)16B
Consumer
DeepSeek V4 Flash284B
Datacenter
DeepSeek V3 (671B MoE)671B
Frontier
DeepSeek V4745B
You are here
DeepSeek V4 Pro1600B
Frontier

Strengths

  • Open-weight benchmark leader (May 2026)
  • 38B active params keep inference practical
  • Strong on coding + math

Weaknesses

  • 745B at any quant requires multi-node cluster
  • Not single-machine deployable

Quantization variants

Each quantization trades model quality for file size and VRAM. Q4_K_M is the most popular starting point.

QuantizationFile sizeVRAM required
AWQ-INT4425.0 GB480 GB

Get the model

HuggingFace

Original weights

huggingface.co/deepseek-ai/DeepSeek-V4

Source repository — direct quantization required.

Hardware that runs this

Cards with enough VRAM for at least one quantization of DeepSeek V4.

NVIDIA GB200 NVL72
13824GB · nvidia
Compare alternatives

Models worth comparing

Same parameter band, plus what's one tier above and below — so you can decide what actually fits your hardware.

Step up
More capable — bigger memory footprint
No verdicted models in the next tier up yet.

Frequently asked

What's the minimum VRAM to run DeepSeek V4?

480GB of VRAM is enough to run DeepSeek V4 at the AWQ-INT4 quantization (file size 425.0 GB). Higher-quality quantizations need more.

Can I use DeepSeek V4 commercially?

Yes — DeepSeek V4 ships under the DeepSeek License, which permits commercial use. Always read the license text before deployment.

What's the context length of DeepSeek V4?

DeepSeek V4 supports a context window of 131,072 tokens (about 131K).

Source: huggingface.co/deepseek-ai/DeepSeek-V4

Reviewed by RunLocalAI Editorial. See our editorial policy for how we research and verify model claims.