Deepseek-V4-Flash-Vision-Exp needs about 104.2 GB of memory at IQ3_XXS (a 103.0 GB file) with an 8K context window, which is more than a 48 GB GPU holds. This repository has 10 quants, from 82.4 GB (IQ1_S) to 161.9 GB (Q8_K_XL). Longer context needs more: at 32K tokens the KV cache alone adds about 2.9 GB.
DeepSeek-V4-Flash-Vision-Exp – 270B multimodal model with vision support
DeepSeek-V4-Flash-Vision-Exp is a 270‑billion‑parameter multimodal model built on the DeepSeek‑4 architecture. It adds visual modules to the DeepSeek‑V4‑Flash base and is trained to handle image inputs alongside text. The model supports a 1,048,576‑token context window and is released under the MIT license.
Good at
- Vision‑text multimodal tasks
- Large context reasoning
- Agent‑style instruction following
Limits
- Requires vision‑enabled llama.cpp (b10766+)
- Large VRAM needed for full precision
- No audio or video input support
Best for
- Image captioning and description
- Multimodal question answering
- Visual‑based instruction following
Who it’s for. Developers building multimodal assistants or agents that need both image and text understanding. The model is suitable for research and production use where a large context window and vision capability are required.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| IQ1_S | 82.4 GB (3 parts) | 83.7 GB | Too big | Get file → | |
| IQ1_M | 86.9 GB (3 parts) | 88.1 GB | Too big | Get file → | |
| IQ2_XXS | 90.7 GB (3 parts) | 91.9 GB | Too big | Get file → | |
| Q2_K_XL | 96.8 GB (3 parts) | 98.1 GB | Too big | Get file → | |
| IQ3_XXS | 103.0 GB (4 parts) | 104.2 GB | Too big | Get file → | |
| IQ3_S | 114.4 GB (4 parts) | 115.6 GB | Too big | Get file → | |
| Q3_K_XL | 128.2 GB (4 parts) | 129.4 GB | Too big | Get file → | |
| IQ4_XS | 136.7 GB (4 parts) | 137.9 GB | Too big | Get file → | |
| Q4_K_XL | 155.1 GB (5 parts) | 156.3 GB | Too big | Get file → | |
| Q8_K_XL | 161.9 GB (5 parts) | 163.1 GB | Too big | Get file → |
For 8‑GB VRAM, use IQ1_S (82.44 GB file, ~83.7 GB memory at 8k context). For 16‑GB VRAM, IQ1_M (86.87 GB file, ~88.1 GB memory). For 32‑GB VRAM, Q2_K_XL (96.83 GB file, ~98.1 GB memory). Higher quants offer better accuracy but need more memory.
What GPU do you need for Deepseek-V4-Flash-Vision-Exp?
These figures add the file size, an fp16 KV cache at an 8K context window (worked out from this model’s own header) 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: no quant fits at an 8K context.
- 24 GB VRAM: no quant fits at an 8K context.
- 32 GB VRAM: no quant fits at an 8K context.
- 48 GB VRAM: no quant fits at an 8K context.
Which Deepseek-V4-Flash-Vision-Exp quant should you choose?
IQ3_XXS (103.0 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 IQ1_S at 82.4 GB, with the quality trade-offs that come with fewer bits. Use it only when memory is very tight. Expect a clear drop in quality, which is usually worse on small models and easier to live with on large ones.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Source
Frequently asked questions about Deepseek-V4-Flash-Vision-Exp
How much VRAM does Deepseek-V4-Flash-Vision-Exp need?
About 104.2 GB at IQ3_XXS with an 8K context window. That is the 103.0 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-Vision-Exp on an 8 GB GPU?
Not at an 8K context. The smallest file needs about 83.7 GB, so you would need a larger GPU or CPU offloading.
What is the file size of Deepseek-V4-Flash-Vision-Exp GGUF?
It depends on the quant: IQ1_S is 82.4 GB, IQ1_M is 86.9 GB, IQ2_XXS is 90.7 GB, Q2_K_XL is 96.8 GB, IQ3_XXS is 103.0 GB, IQ3_S is 114.4 GB and 4 more.
What context length does Deepseek-V4-Flash-Vision-Exp support?
The GGUF header lists a trained context length of 1,048,576 tokens (1024K). The KV cache grows with the context you use: about 0.7 GB at 8K and 2.9 GB at 32K.
Where can I download Deepseek-V4-Flash-Vision-Exp GGUF?
The files are on Hugging Face in the unsloth/DeepSeek-V4-Flash-Vision-Exp-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-Vision-Exp 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.
Can I run DeepSeek-V4-Flash-Vision-Exp on 8 GB VRAM?
Only with the smallest quant IQ1_S, which still requires ~83.7 GB memory at 8k context; 8 GB VRAM is insufficient.
Does the model support video input?
No, the model only handles image inputs; video is not supported.
What license applies to this model?
The model is released under the MIT license.