DeepSeek V4 Flash 284B
DeepSeek V4 Flash 284B needs roughly 179.9 GB VRAM at Q4_K_M quantization (637.0 GB at FP16). 14 GPUs we track can run it fully in VRAM at 8k context.
14 GPUs run this natively · 1 with CPU offload
DeepSeek V4 Flash 284B is a Mixture of Experts (MoE) large language model with 284B total parameters but only 13B active per token developed by DeepSeek. Released in April 2026, it supports text and vision and video inputs with a 128K context window, released under the MIT license, allowing commercial use. Lighter V4 sibling — 284B total / 13B active, same 1M context and hybrid CSA/HCA attention as Pro.
To run DeepSeek V4 Flash 284B locally, you need approximately 179.9 GB of VRAM at Q4_K_M quantization with 8k context. 14 of the GPUs we track can run it fully in VRAM, with a further 1 able to offload to system RAM. At Q4_K_M it requires 179.9 GB — more than any single consumer GPU. A multi-GPU server or 80 GB datacenter GPU is required. At Q8_K_M (318.9 GB), you get near-FP16 quality while still fitting on large server or multi-GPU setups. FP16 requires 637.0 GB, limiting it to datacenter-class hardware with 80 GB+ VRAM. As a MoE model, inference speed depends on active parameters (13B) rather than total size, so once it fits in VRAM it runs noticeably faster than dense models of the same parameter count.
Its MMLU-Pro score of 86.3 places it among the strongest open-weight models available. The license allows commercial use.
VRAM at each quantization
Assumes 8k context. KV cache grows linearly with context length.
| Quant | Weights | KV cache | Total |
|---|---|---|---|
| FP32 | 1136.0 GB | 0.72 GB | 1273.1 GB |
| BF16 | 568.0 GB | 0.72 GB | 637.0 GB |
| FP16 | 568.0 GB | 0.72 GB | 637.0 GB |
| Q8_0 | 284.0 GB | 0.72 GB | 318.9 GB |
| Q6_K | 232.9 GB | 0.72 GB | 261.6 GB |
| Q5_K_M | 182.9 GB | 0.72 GB | 205.7 GB |
| Q4_K_M | 159.9 GB | 0.72 GB | 179.9 GB |
| Q3_K_M | 122.1 GB | 0.72 GB | 137.6 GB |
| Q2_Krec | 93.4 GB | 0.72 GB | 105.5 GB |
| NVFP4cuda | 142.0 GB | 0.72 GB | 159.8 GB |
KV cache shown at 8k context (FP16). NVFP4 requires a CUDA GPU. Enable TurboQuant in the calculator to see reduced KV cache estimates.
Benchmarks
GPUs that run DeepSeek V4 Flash 284B natively (14)
- NVIDIA DGX Spark (128GB)Q2_K · 70.2 t/s
- AMD Instinct MI300XQ4_K_M · 796.6 t/s
- AMD Strix Halo (128GB)Q2_K · 65.8 t/s
- Apple M5 Max (128GB)Q2_K · 157.9 t/s
- Apple M4 Ultra (384GB)Q8_0 · 92.4 t/s
- Apple M4 Ultra (192GB)Q4_K_M · 164.1 t/s
- Apple M4 Max (128GB)Q2_K · 140.4 t/s
- Apple M3 Ultra (512GB)Q8_0 · 69.3 t/s
- Apple M3 Ultra (256GB)Q5_K_M · 107.6 t/s
- Apple M3 Max (128GB)Q2_K · 102.9 t/s
- Apple M2 Ultra (384GB)Q8_0 · 67.7 t/s
- Apple M2 Ultra (192GB)Q4_K_M · 120.2 t/s
- Apple M1 Ultra (128GB)Q2_K · 205.8 t/s
- Intel Data Center GPU Max 1550Q2_K · 842.6 t/s
Plus 1 GPUs that run it with CPU offload (slower)
- NVIDIA RTX Pro 6000Q2_K · 78.6 t/s
Notes
Lighter V4 sibling — 284B total / 13B active, same 1M context and hybrid CSA/HCA attention as Pro.
Frequently asked questions
- What are the VRAM requirements for DeepSeek V4 Flash 284B?
- DeepSeek V4 Flash 284B requires approximately 179.9 GB of VRAM at Q4_K_M quantization, 318.9 GB at Q8, and 637.0 GB at FP16. These numbers assume 8k context window; VRAM scales linearly with context length due to the KV cache.
- How many parameters does DeepSeek V4 Flash 284B have?
- DeepSeek V4 Flash 284B has 284 billion total parameters, but only 13 billion are active per token thanks to its Mixture of Experts (MoE) architecture. This makes inference significantly faster than the total parameter count suggests.
- How capable is DeepSeek V4 Flash 284B?
- DeepSeek V4 Flash 284B achieves an MMLU-Pro score of 86.3, placing it among the most capable open-weight models available — competitive with frontier systems on general knowledge and reasoning.
- Can DeepSeek V4 Flash 284B run on a 16 GB GPU?
- No. At Q4_K_M, DeepSeek V4 Flash 284B needs 179.9 GB of VRAM — more than 16 GB. You will need a multi-GPU server.
- Can DeepSeek V4 Flash 284B run on a 24 GB GPU?
- No. Even at Q4_K_M, DeepSeek V4 Flash 284B needs 179.9 GB. Consider a multi-GPU server with 80 GB+ total VRAM.
- What is the smallest quantization for DeepSeek V4 Flash 284B that fits in 24 GB of VRAM?
- DeepSeek V4 Flash 284B cannot fit in 24 GB of VRAM at any standard quantization level. The minimum needed is 105.5 GB at Q2_K.
- What GPU do I need to run DeepSeek V4 Flash 284B locally?
- You need a multi-GPU server. At Q4_K_M, DeepSeek V4 Flash 284B needs 179.9 GB VRAM, more than any single consumer GPU. Consider 2–4× H100 or A100 GPUs.