Skip to content

DeepSeek V4.1 Flash on 4× RTX PRO 6000 Blackwell: VRAM, speed and fit

DeepSeek V4.1 Flash at 3-bit runs on the 4× RTX PRO 6000 Blackwell with 1M of context at about 111 tokens a second.

Text generation · Chat · Ollama / LM Studio

Runs comfortably · on 4× RTX PRO 6000 Blackwell · 3-bit (Q3_K_M) · Ollama / LM Studio · Chat

GPU memory
353 GB
of 384 GB
Context
1M
auto
Speed
111 tok/s
89 when full
First reply
0.4 s
2K prompt

Where the memory goes

Model
347 GB
Context
2.6 GB
Overhead
2.8 GB
OS
0.0 GB
Free
31.2 GB
Memory
353 GB of 384 GB
Context
1M tokens · auto
Speed
111 tok/s · 89 when full
First reply
0.4 s for a 2K prompt
Running cost
$33/mo at home · $574/mo on a cloud 8× A40

Validation: Validated

What DeepSeek V4.1 Flash on 4× RTX PRO 6000 Blackwell was checked against
CheckTargetResultStatus
Fits in GPU memory384 GB usable353 GB neededPass
Headroom5% free31.2 GB free (8%)Pass
Context for chat16K or more1MPass
Writing speed15 tok/s or more111 tok/sPass
First replyunder 3 s, 2K prompt0.4 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 V4.1 Flash
Parameters
763B, 16B active
Precision
3-bit (Q3_K_M)
Context
1M, auto
Chats at once
1
Use
Chat
Runs in
Ollama / LM Studio
Hardware
4× RTX PRO 6000 Blackwell
GPU can use
384 GB of 384 GB

Running cost

At home
$33/mo
1520 W at the wall
Renting instead
$574/mo
8× A40 at $4.72/h

4 h a day, power at $0.18 per kWh, cloud at RunPod on-demand prices as of Nov 22, 2026.

How to run it

llama.cpp
llama-server -hf unsloth/DeepSeek-V4.1-Flash-1017-GGUF
vLLM
vllm serve deepseek-ai/DeepSeek-V4.1-Flash
MLX (Mac)
mlx_lm.chat --model mlx-community/DeepSeek-V4.1-Flash-MLX-4bit

DeepSeek V4.1 Flash on similar machines

Other models that run on 4× RTX PRO 6000 Blackwell

Questions

Can DeepSeek V4.1 Flash run on 4× RTX PRO 6000 Blackwell?

DeepSeek V4.1 Flash at 3-bit runs on the 4× RTX PRO 6000 Blackwell with 1M of context at about 111 tokens a second.

Report VR-0LYMEC0, calculated Nov 22, 2026 from published specs, not benchmarked.