Can I Run / DeepSeek R1 0528 / on NVIDIA DGX Station (Blackwell Ultra)

Can I Run DeepSeek R1 0528 on a NVIDIA DGX Station (Blackwell Ultra)?

Yes

Runs at Q8_0 — near-lossless quality.

Model size
685B
GPU memory
784GB
Smallest quant
Q4_K_M
Best fit
Q8_0

4 quantizations fit your 784GB

QuantMin VRAMRecommendedFile sizeHeadroom
Q8_0BEST728.8 GB730.3 GB727.8 GB+55.2 GB
Q6_K565.3 GB566.8 GB564.3 GB+218.7 GB
Q5_K_M487.4 GB488.9 GB486.4 GB+296.6 GB
Q4_K_M416.3 GB417.8 GB415.3 GB+367.7 GB

Try it in the cloud first

Don't want to download DeepSeek R1 0528 just to try it? Use a hosted API or rent a GPU by the second.

Affiliate links — we earn a commission at no cost to you.

Advertisement
Full model details
DeepSeek R1 0528

All quant variants, benchmark scores, and use-case tags.

Best models for this GPU
NVIDIA DGX Station (Blackwell Ultra)

Top-ranked open-source models that fit in 784GB.

FAQ

Can the NVIDIA DGX Station (Blackwell Ultra) run DeepSeek R1 0528?

Yes. The NVIDIA DGX Station (Blackwell Ultra)'s 784GB of unified memory is enough to run DeepSeek R1 0528 at Q8_0 quantization (728.8GB required).

What's the best quantization to use?

Q8_0 is the highest-precision quantization that fits in your 784GB. It uses about 728.8GB of memory and 730.3GB recommended for comfortable inference.

What if I need more headroom for context length?

KV cache memory grows with context length. The numbers above assume a baseline 2K-4K context. For long-context use (32K+), add another 2-6GB depending on the model architecture.