DE

DeepSeek V4.1 Flash MXFP4

DeepSeekGGUF by kernelpoolUpdated 2026-10-22
Download on Hugging Face
Parameters485M
Context—
Layers32
Architecturedeepseek4-vision
Smallest quant1.0 GB
Quick answer

DeepSeek V4.1 Flash MXFP4 needs about 2.5 GB of memory at its main file (a 1.0 GB file) with an 8K context window, so it fits on a 4 GB GPU. This repository has 2 quants, from 1.0 GB to 8.6 GB.

DeepSeek V4.1 Flash MXFP4 – multimodal LLM with compressed KV cache

DeepSeek V4.1 Flash MXFP4 is a multimodal model based on the DeepSeek‑4 architecture. It supports image and text input and generates text autoregressively. The model uses compressed sparse attention and a Causal Encoder‑Decoder design to reduce KV cache size.

Good at

  • Multimodal text and image generation
  • Efficient KV cache compression
  • Large context handling up to one million tokens

Limits

  • Requires DS4 layout, not llama.cpp
  • Large memory footprint – 10.2 GB at 8 k context for full file
  • No explicit language support listed

Best for

  • Generating image‑caption pairs
  • Chatting with image context
  • Research on multimodal LLMs

Who it’s for. Developers and researchers needing a multimodal LLM with efficient KV cache. The model is suitable for environments that can run DS4 and have sufficient VRAM to load the 8.63 GB file.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
—1.0 GB 2.5 GB
Fits Get file →
—8.6 GB 10.2 GB
Too big Get file →

For 8 GB VRAM, use the 0.97 GB file which estimates 2.5 GB memory at 8 k context. For larger VRAM budgets, the 8.63 GB file uses about 10.2 GB memory at 8 k context. Both files are quantized as "unknown" and require DS4 support.

What GPU do you need for DeepSeek V4.1 Flash MXFP4?

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: the best quant that fits is the main file (1.0 GB file, about 2.5 GB in use).
  • 6 GB VRAM: the best quant that fits is the main file (1.0 GB file, about 2.5 GB in use).
  • 8 GB VRAM: the best quant that fits is the main file (1.0 GB file, about 2.5 GB in use).
  • 12 GB VRAM: the best quant that fits is the main file (8.6 GB file, about 10.2 GB in use).

Specifications

Architecture from the GGUF header

Architecturedeepseek4-vision
Parameters485M
Layers32
KV cache (fp16)128 KB per token

Source

Quantized bykernelpool
FamilyDeepSeek
Licensemit
Downloads (HF)932
Last updated2026-10-22

Frequently asked questions about DeepSeek V4.1 Flash MXFP4

How much VRAM does DeepSeek V4.1 Flash MXFP4 need?

About 2.5 GB at its main file with an 8K context window. That is the 1.0 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 V4.1 Flash MXFP4 on an 8 GB GPU?

Yes, with the smaller files: the 1.0 GB file needs about 2.5 GB at 8K context. Larger quants in this repository do not fit in 8 GB.

What is the file size of DeepSeek V4.1 Flash MXFP4 GGUF?

It depends on the quant: file is 1.0 GB, file is 8.6 GB.

Where can I download DeepSeek V4.1 Flash MXFP4 GGUF?

The files are on Hugging Face in the kernelpool/DeepSeek-V4.1-Flash-MXFP4-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 V4.1 Flash MXFP4 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 this model on 8 GB VRAM?

Yes, the 0.97 GB file uses an estimated 2.5 GB memory at 8 k context, fitting within 8 GB VRAM.

Does this model support vision?

Yes, it includes a vision encoder and can process images with the --vision flag.

What quantization is used?

The GGUF files are marked as "unknown" quant, and no specific quant name is provided.

Other GGUFs of DeepSeek-V4.1-Flash

Similar models

Browse all →

More from kernelpool

All by kernelpool →

More DeepSeek models

All DeepSeek →

Other tiny models

All tiny models →