DE

DeepSeek V4 Flash Vision Exp

DeepSeekGGUF by ggml-orgUpdated 2026-10-19
Download on Hugging Face
Parameters284B
Context—
Layers—
Architecturedflash
Smallest quant10.8 GB
Quick answer

DeepSeek V4 Flash Vision Exp needs about 110.9 GB of memory at Q2_K_S (a 98.6 GB file) with an 8K context window, which is more than a 48 GB GPU holds. This repository has 5 quants, from 10.8 GB (MXFP4) to 155.0 GB (MXFP4).

DeepSeek V4 Flash Vision Exp – multimodal LLM with vision encoder

DeepSeek V4 Flash Vision Exp is a multimodal model built on the DeepSeek V4 Flash architecture. It adds visual modules and continues training to provide visual understanding while maintaining text‑only performance. The model is released under the MIT license.

Good at

  • Vision‑text multimodal tasks
  • Text‑only agent tasks
  • Experimental multimodal agent capabilities

Limits

  • No explicit vision support in benchmarks
  • Large memory requirement for full precision
  • No mention of training context tokens

Best for

  • Multimodal content generation
  • Vision‑enabled chatbots
  • Experimental research on multimodal agents

Who it’s for. Researchers and developers interested in multimodal LLMs who need a model that combines text and vision. The model is suitable for experimentation and prototyping in environments that support GGUF formats.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
MXFP410.8 GB 12.6 GB
Too big Get file →
BF1610.9 GB 12.7 GB
Too big Get file →
Q2_K_S98.6 GB (2 parts) 110.9 GB
Too big Get file →
Q2_K117.3 GB (2 parts) 131.9 GB
Too big Get file →
MXFP4155.0 GB (2 parts) 174.1 GB
Too big Get file →

For 8 GB VRAM, the MXFP4 quant at 10.83 GB file size uses about 12.6 GB memory at 8 k context, which is beyond 8 GB. The BF16 quant at 10.9 GB file size uses about 12.7 GB memory, also above 8 GB. For 16 GB VRAM, the MXFP4 quant is suitable, using roughly 12.6 GB memory. Larger models such as Q2_K_S or Q2_K require over 110 GB memory and are not recommended for typical consumer GPUs.

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 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 (10.9 GB file, about 12.7 GB in use).
  • 24 GB VRAM: the best quant that fits is BF16 (10.9 GB file, about 12.7 GB in use).
  • 32 GB VRAM: the best quant that fits is BF16 (10.9 GB file, about 12.7 GB in use).
  • 48 GB VRAM: the best quant that fits is BF16 (10.9 GB file, about 12.7 GB in use).

Which DeepSeek V4 Flash Vision Exp quant should you choose?

Q2_K_S (98.6 GB) is the middle option in this repository. 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.

If memory is tight, the smallest file is MXFP4 at 10.8 GB, with the quality trade-offs that come with fewer bits. Lower bit counts mean smaller files and more quality loss.

New to quant names? Read GGUF quantization explained.

Specifications

Architecture from the GGUF header

Architecturedflash
Parameters284B

Source

Quantized byggml-org
FamilyDeepSeek
Licensemit
Downloads (HF)2,615
Last updated2026-10-19

Frequently asked questions about DeepSeek V4 Flash Vision Exp

How much VRAM does DeepSeek V4 Flash Vision Exp need?

About 110.9 GB at Q2_K_S with an 8K context window. That is the 98.6 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 12.6 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: MXFP4 is 10.8 GB, BF16 is 10.9 GB, Q2_K_S is 98.6 GB, Q2_K is 117.3 GB, MXFP4 is 155.0 GB.

Where can I download DeepSeek V4 Flash Vision Exp GGUF?

The files are on Hugging Face in the ggml-org/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?

The smallest quant requires about 12.6 GB memory at 8 k context, so 8 GB VRAM is insufficient.

What license does this model use?

The model is released under the MIT license.

Does the model support vision tasks?

Yes, it includes a vision encoder and is designed for multimodal tasks.

Other GGUFs of DeepSeek-V4-Flash-Vision-Exp

Similar models

Browse all →

More from ggml-org

All by ggml-org →

More DeepSeek models

All DeepSeek →

Other large models

All large models →