CanItRun Logocanitrun.

DeepSeek V4 Flash 284B

DeepSeek V4 Flash 284B needs roughly 159.8GB VRAM at Q4_K_M quantization (637.0GB at FP16). 10 GPUs we track can run it fully in VRAM at 8k context.

DeepSeek284B params13B active (MoE)1024k contextMITCommercial use ok

VRAM at each quantization

Assumes 8k context. KV cache grows linearly with context length.

QuantWeightsKV cacheTotal
FP16568.0 GB0.72 GB637.0 GB
Q8284.0 GB0.72 GB318.9 GB
Q6_K213.0 GB0.72 GB239.4 GB
Q5_K_M177.5 GB0.72 GB199.6 GB
Q4_K_M142.0 GB0.72 GB159.8 GB
Q3_K_M113.6 GB0.72 GB128.0 GB
Q2_K85.2 GB0.72 GB96.2 GB

Benchmarks

Benchmarks for this model are not yet available on the Open LLM Leaderboard v2. This is common for recently released models. Check back soon.

GPUs that run DeepSeek V4 Flash 284B natively (10)

Plus 2 GPUs that run it with CPU offload (slower)

Notes

Lighter V4 sibling — 284B total / 13B active, same 1M context and hybrid CSA/HCA attention as Pro.

Hugging Face ↗Released 2026-04-24