Question 1

How much VRAM does DeepSeek-R1-0528-NVFP4-v2 need?

Accepted Answer

DeepSeek-R1-0528-NVFP4-v2 (393.6B) requires approximately 879.8379152297973 GB of VRAM at full precision (FP16), 439.9 GB at 8-bit quantization (Q8), and 220 GB at 4-bit quantization (Q4). The actual VRAM usage may be slightly higher due to KV cache and runtime overhead.

Question 2

What GPUs can run DeepSeek-R1-0528-NVFP4-v2?

Accepted Answer

We currently have no compatible GPUs indexed for DeepSeek-R1-0528-NVFP4-v2. This may be because the model's VRAM requirements exceed available GPUs in our database, or VRAM data hasn't been computed yet.

Question 3

What is the cheapest GPU to run DeepSeek-R1-0528-NVFP4-v2?

Accepted Answer

We don't currently have cloud pricing data for GPUs compatible with DeepSeek-R1-0528-NVFP4-v2. Cloud pricing data is updated regularly — check back soon.

Question 4

Can DeepSeek-R1-0528-NVFP4-v2 run on consumer hardware?

Accepted Answer

DeepSeek-R1-0528-NVFP4-v2 requires approximately 879.8379152297973 GB of VRAM at FP16, and even with 4-bit quantization (Q4) it needs about 220.0 GB. This exceeds the 24 GB available on the highest-end consumer GPU (RTX 4090), so DeepSeek-R1-0528-NVFP4-v2 requires datacenter or cloud GPUs.

Question 5

What quantization options are available for DeepSeek-R1-0528-NVFP4-v2?

Accepted Answer

DeepSeek-R1-0528-NVFP4-v2 supports the following quantization levels: FP16 (full precision): ~879.8379152297973 GB, Q8 (8-bit): ~439.9 GB, Q6 (6-bit): ~329.9 GB, Q4 (4-bit): ~220 GB. Lower quantization reduces VRAM usage at the cost of some quality loss. Q8 offers a good balance between quality and memory savings, while Q4 provides maximum compression. As a Mixture of Experts (MoE) model with 256 experts (8 active), DeepSeek-R1-0528-NVFP4-v2 has a larger total weight size but only activates a subset of parameters per token, which can affect effective memory usage during inference.

DeepSeek-R1-0528-NVFP4-v2

VRAM Requirements

Compatible GPUs

DeepSeek-R1-0528-NVFP4-v2 FAQ

VRAM requirements for DeepSeek-R1-0528-NVFP4-v2 at different quantization levels
Quantization	VRAM Required
FP16	879.8 GB
Q8	439.9 GB
Q6	329.9 GB
Q4	220 GB