Skip to content

DeepSeek V4 Flash on DGX Spark: VRAM, speed and fit

DeepSeek V4 Flash at 2-bit runs on the DGX Spark 128 GB with 1M of context at about 25 tokens a second.

Text generation · Chat · Ollama / LM Studio

Runs comfortably · on DGX Spark 128 GB · 2-bit (Q2_K) · Ollama / LM Studio · Chat

GPU memory
114 GB
of 128 GB
Context
1M
auto
Speed
25 tok/s
14 when full
First reply
2.8 s
2K prompt

Where the memory goes

Model
99.2 GB
Context
6.1 GB
Overhead
637 MB
OS
7.7 GB
Free
14.4 GB
Memory
114 GB of 128 GB
Context
1M tokens · auto
Speed
25 tok/s · 14 when full
First reply
2.8 s for a 2K prompt
Running cost
$3.15/mo at home · $287/mo on a cloud 4× A40

Validation: Validated

What DeepSeek V4 Flash on DGX Spark was checked against
CheckTargetResultStatus
Fits in GPU memory120 GB usable106 GB neededPass
Headroom5% free14.4 GB free (12%)Pass
Context for chat16K or more1MPass
Writing speed15 tok/s or more25 tok/sPass
First replyunder 3 s, 2K prompt2.8 sPass

Targets: a reply you can read along with, a first reply the use case tolerates, and 5% of memory left free. Calculated from published specs, not benchmarked.

The setup

Model
DeepSeek V4 Flash
Parameters
284B, 13B active
Precision
2-bit (Q2_K)
Context
1M, auto
Chats at once
1
Use
Chat
Runs in
Ollama / LM Studio
Hardware
DGX Spark 128 GB
GPU can use
120 GB of 128 GB

Running cost

At home
$3.15/mo
144 W at the wall
Renting instead
$287/mo
4× A40 at $2.36/h

4 h a day, power at $0.18 per kWh, cloud at RunPod on-demand prices as of Nov 21, 2026.

How to run it

llama.cpp
llama-server -hf unsloth/DeepSeek-V4-Flash-0731-GGUF
vLLM
vllm serve deepseek-ai/DeepSeek-V4-Flash-0731
MLX (Mac)
mlx_lm.chat --model mlx-community/DeepSeek-V4-Flash-0731-2.4bit-mixed

DeepSeek V4 Flash on similar machines

All machines: DeepSeek V4 Flash VRAM requirements. Everything that runs here: what can the DGX Spark run?

Other models that run on DGX Spark

Questions

Can DeepSeek V4 Flash run on DGX Spark?

DeepSeek V4 Flash at 2-bit runs on the DGX Spark 128 GB with 1M of context at about 25 tokens a second.

Report VR-11AQIUZ, calculated Nov 22, 2026 from published specs, not benchmarked.