ToolSite DeepSeek V4 Flash VRAM requirements
DeepSeek V4 Flash has 290.9B parameters, of which about 13B are used per token; all 256 experts still have to be in memory. With an 8K-token context and one request it needs about 182 GB of GPU memory at Q4_K_M, 299 GB at FP8 and 597 GB at FP16/BF16. The published weights take 149 GB (4.4 bits/weight). No single GPU in the table holds it at Q4_K_M; it needs a multi-GPU setup.
Open DeepSeek V4 Flash in the calculator
VRAM by quantization
Weights plus the KV cache for 8,192 tokens in FP16 and the runtime overhead (0.5 GB plus 10%). Each row opens the calculator with that setting. What the GGUF names mean.
| Precision | Weights | Total | Smallest setup |
|---|---|---|---|
| As published (4.4 bits/weight) | 149 GB | 165 GB | B200 |
| FP16 / BF16 | 542 GB | 597 GB | 4 × B200 |
| FP8 / INT8 | 271 GB | 299 GB | 2 × B200 |
| INT4 (AWQ / GPTQ) | 144 GB | 160 GB | B200 |
| GGUF Q8_0 | 288 GB | 318 GB | 2 × B200 |
| GGUF Q6_K | 222 GB | 246 GB | 2 × H200 |
| GGUF Q5_K_M | 192 GB | 212 GB | 2 × H200 |
| GGUF Q4_K_M | 164 GB | 182 GB | 2 × H200 |
| GGUF Q3_K_M | 132 GB | 147 GB | B200 |
| GGUF Q2_K | 113 GB | 126 GB | H200 |
KV cache at long context
All 43 layers use full attention. Each extra token of context adds 86 KB of FP16 cache per request. This model shares and compresses its KV cache across layers, which the calculator does not model, so the KV cache figure is an upper bound. How the KV cache works.
| Context | KV cache, FP16 | KV cache, FP8 | Total at Q4_K_M |
|---|---|---|---|
| 4K tokens | 344 MB | 172 MB | 181 GB |
| 32K tokens | 2.69 GB | 1.34 GB | 184 GB |
| 128K tokens | 10.8 GB | 5.38 GB | 193 GB |
| 1M tokens | 86.0 GB | 43.0 GB | 275 GB |
Which GPUs can run DeepSeek V4 Flash
With 8,192 tokens of context. Several GPUs means one tensor-parallel group of 2, 4 or 8 cards.
| GPU | Memory | Q4_K_M | FP8 |
|---|---|---|---|
| RTX 3060 | 12 GB | Needs more than 8 | Needs more than 8 |
| RTX 4060 Ti 16GB | 16 GB | Needs more than 8 | Needs more than 8 |
| RTX 3090 / 4090 | 24 GB | Needs 8 | Needs more than 8 |
| RTX 5090 | 32 GB | Needs 8 | Needs more than 8 |
| A100 40GB | 40 GB | Needs 8 | Needs 8 |
| Mac, 64 GB unified memory about 75% of it is usable by the GPU by default | 48 GB | Needs 4 | Needs 8 |
| L40S / RTX 6000 Ada | 48 GB | Needs 4 | Needs 8 |
| A100 / H100 80GB | 80 GB | Needs 4 | Needs 4 |
| Mac, 128 GB unified memory about 75% of it is usable by the GPU by default | 96 GB | Needs 2 | Needs 4 |
| H200 | 141 GB | Needs 2 | Needs 4 |
| B200 | 180 GB | Needs 2 | Needs 2 |
How fast DeepSeek V4 Flash writes
Tokens per second for one request with 8,192 tokens of context, estimated from memory bandwidth and the parameters read per token. A dash means it does not fit on one card. Try other settings in the speed calculator.
| Hardware | Bandwidth | Q4_K_M | FP8 |
|---|---|---|---|
| RTX 3060 12GB | 360 GB/s | — | — |
| RTX 4090 | 1,008 GB/s | — | — |
| RTX 5090 | 1,792 GB/s | — | — |
| M4 Max Mac (128 GB) | 546 GB/s | — | — |
| M3 Ultra Mac Studio (512 GB) | 819 GB/s | 27–47 | 17–29 |
| H100 SXM | 3,350 GB/s | — | — |
| H200 | 4,800 GB/s | — | — |
Longest context on one GPU
How many tokens of context fit on a single card with one request and an FP16 KV cache. "Full" means the model's whole context window fits.
| GPU | Memory | Q4_K_M | Q8_0 | FP8 |
|---|---|---|---|---|
| RTX 3060 | 12 GB | No | No | No |
| RTX 4060 Ti 16GB | 16 GB | No | No | No |
| RTX 3090 / 4090 | 24 GB | No | No | No |
| RTX 5090 | 32 GB | No | No | No |
| A100 40GB | 40 GB | No | No | No |
| Mac, 64 GB unified memory | 48 GB | No | No | No |
| L40S / RTX 6000 Ada | 48 GB | No | No | No |
| A100 / H100 80GB | 80 GB | No | No | No |
| Mac, 128 GB unified memory | 96 GB | No | No | No |
| H200 | 141 GB | No | No | No |
| B200 | 180 GB | No | No | No |
Model details
- Parameters
- 290.9B (290,944,616,402)
- Experts
- 256 routed experts, all loaded
- Active per token
- 13B
- Layers
- All 43 layers use full attention
- Attention cache
- 1 KV heads × 512
- Context length
- 1,048,576 tokens
- Published weights
- 149 GB (4.4 bits/weight)
- On Hugging Face
- deepseek-ai/DeepSeek-V4-Flash
Other models
- DeepSeek V4 Flash 0731
- Qwen3.6 35B-A3B
- Qwen3.6 27B
- Qwen3.8 27B
- Qwen3.8 Flash Next
- Qwen3.8 2.4T-A95B
- Qwen3.5 9B
- Qwen3.5 122B-A10B
- Qwen3-Coder-Next
- DeepSeek V4.1 Flash
- DeepSeek V4 Pro
- DeepSeek V3.2
- GLM-5.3
- GLM-5.3 Flash
- GLM-5.2
- GLM-4.7 Flash
- Gemma 4 31B
- Gemma 4 26B-A4B
- Gemma 4 12B
- Gemma 4 E4B
- Kimi K3
- MiniMax M3
- MiniMax M2.7
- MiMo V2.6 Flash
- MiMo V2.6 Pro
- Mistral Medium 3.5 128B
- Nemotron 3 Nano 4B
- Nemotron 3 Nano 30B-A3B
- Nemotron 3 Super 120B-A12B
- Xing 4.0 29B-A4B
- Muse Glimmer 30B
- MiniCPM5 2B
- Llama 3.1 8B
- Llama 3.1 70B
- Qwen3 8B
- Qwen3 30B-A3B
- gpt-oss-20b
- gpt-oss-120b
- DeepSeek V3 / R1
Numbers read from the model files on Hugging Face on .