DeepSeek V4 Flash 0731 needs about 117.2 GB of memory at IQ3_XXS (a 104.2 GB file) with an 8K context window, which is more than a 48 GB GPU holds. This repository has 15 quants, from 10.9 GB (Q8_0) to 161.9 GB (Q8_K_XL).
DeepSeek V4 Flash 0731 is a 284B-parameter DeepSeek model in GGUF format, quantized by Stk88. Its trained context length is 1024K tokens. The repository has 15 quant files, from 10.9 GB (Q8_0) to 161.9 GB (Q8_K_XL).
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| Q8_0 | 10.9 GB | 12.7 GB | Too big | Get file → | |
| BF16 | 11.3 GB | 13.2 GB | Too big | Get file → | |
| IQ1_S | 82.5 GB (3 parts) | 92.9 GB | Too big | Get file → | |
| IQ1_M | 86.9 GB (3 parts) | 97.8 GB | Too big | Get file → | |
| IQ2_XXS | 90.9 GB (3 parts) | 102.3 GB | Too big | Get file → | |
| IQ2_M | 90.9 GB (3 parts) | 102.3 GB | Too big | Get file → | |
| Q2_K_XL | 96.8 GB (3 parts) | 109.0 GB | Too big | Get file → | |
| IQ3_XXS | 104.2 GB (4 parts) | 117.2 GB | Too big | Get file → | |
| IQ3_S | 116.1 GB (4 parts) | 130.5 GB | Too big | Get file → | |
| Q3_K_M | 128.1 GB (4 parts) | 143.9 GB | Too big | Get file → | |
| Q3_K_XL | 128.2 GB (4 parts) | 144.1 GB | Too big | Get file → | |
| IQ4_NL | 136.7 GB (4 parts) | 153.6 GB | Too big | Get file → | |
| IQ4_XS | 136.7 GB (4 parts) | 153.6 GB | Too big | Get file → | |
| Q4_K_XL | 155.1 GB (5 parts) | 174.2 GB | Too big | Get file → | |
| Q8_K_XL | 161.9 GB (5 parts) | 181.8 GB | Too big | Get file → |
What GPU do you need for DeepSeek V4 Flash 0731?
These figures add the file size, an fp16 KV cache at an 8K context window and about 0.5 GB for the runtime. They are estimates, so leave a little headroom, and expect to use less context or a smaller quant if you are right at the limit.
- 4 GB VRAM: no quant fits at an 8K context.
- 6 GB VRAM: no quant fits at an 8K context.
- 8 GB VRAM: no quant fits at an 8K context.
- 12 GB VRAM: no quant fits at an 8K context.
- 16 GB VRAM: the best quant that fits is BF16 (11.3 GB file, about 13.2 GB in use).
- 24 GB VRAM: the best quant that fits is BF16 (11.3 GB file, about 13.2 GB in use).
- 32 GB VRAM: the best quant that fits is BF16 (11.3 GB file, about 13.2 GB in use).
- 48 GB VRAM: the best quant that fits is BF16 (11.3 GB file, about 13.2 GB in use).
Which DeepSeek V4 Flash 0731 quant should you choose?
IQ3_XXS (104.2 GB) is the middle option in this repository. A reasonable choice when a 4-bit file will not fit, mostly on larger models. Quality loss is noticeable but often acceptable.
If memory is tight, the smallest file is Q8_0 at 10.9 GB, with the quality trade-offs that come with fewer bits. Close to lossless. It costs roughly twice the memory of a 4-bit file, so it suits cases where you have plenty of VRAM.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Source
Frequently asked questions about DeepSeek V4 Flash 0731
How much VRAM does DeepSeek V4 Flash 0731 need?
About 117.2 GB at IQ3_XXS with an 8K context window. That is the 104.2 GB file, plus the KV cache and roughly 0.5 GB of runtime overhead. Smaller quants need less and longer context needs more.
Can I run DeepSeek V4 Flash 0731 on an 8 GB GPU?
Not at an 8K context. The smallest file needs about 12.7 GB, so you would need a larger GPU or CPU offloading.
What is the file size of DeepSeek V4 Flash 0731 GGUF?
It depends on the quant: Q8_0 is 10.9 GB, BF16 is 11.3 GB, IQ1_S is 82.5 GB, IQ1_M is 86.9 GB, IQ2_XXS is 90.9 GB, IQ2_M is 90.9 GB and 9 more.
What context length does DeepSeek V4 Flash 0731 support?
The GGUF header lists a trained context length of 1,048,576 tokens (1024K).
Where can I download DeepSeek V4 Flash 0731 GGUF?
The files are on Hugging Face in the Stk88/DeepSeek-V4-Flash-0731-GGUF repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.
What license does DeepSeek V4 Flash 0731 use?
Hugging Face lists the license as mit. Licenses can differ between the quantized files and the original model, so check the repository and the base model before commercial use.