DE

DeepSeek V4 Flash 0731 IK

DeepSeekGGUF by KeinNiemandUpdated 2026-09-14
Download on Hugging Face
Parameters284B
Context—
Layers—
Architecturedeepseek4
Smallest quant65.4 GB
Quick answer

DeepSeek V4 Flash 0731 IK needs about 119.6 GB of memory at its main file (a 106.3 GB file) with an 8K context window, which is more than a 48 GB GPU holds. This repository has 9 quants, from 65.4 GB to 161.9 GB (MXFP4).

DeepSeek V4 Flash 0731 IK – 284 billion‑parameter LLM with speculative decoding

DeepSeek V4 Flash 0731 IK is a 284 billion‑parameter model built on the deepseek4 architecture. It supports a large context window and includes a speculative decoding module. The model is released under the MIT license and is available in GGUF format.

Good at

  • Large context handling
  • Speculative decoding
  • Agentic reasoning

Limits

  • No vision support
  • Requires recent ik_llama.cpp
  • Large memory footprint

Best for

  • Chat applications needing long context
  • Agentic coding tasks
  • Research on speculative decoding

Who it’s for. Engineers and researchers who need a high‑parameter model with a large context window and speculative decoding. The model is suitable for developers using llama.cpp or compatible backends that support IK tensor types.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
—65.4 GB 73.8 GB
Too big Get file →
—73.2 GB 82.5 GB
Too big Get file →
—93.0 GB 104.7 GB
Too big Get file →
—93.4 GB 105.2 GB
Too big Get file →
—106.3 GB 119.6 GB
Too big Get file →
—122.0 GB 137.2 GB
Too big Get file →
—132.8 GB 149.2 GB
Too big Get file →
—149.2 GB 167.6 GB
Too big Get file →
MXFP4161.9 GB 181.8 GB
Too big Get file →

For 8 GB VRAM use the MXFP4 quantized file (161.87 GB). For 16 GB VRAM use the 149.24 GB file. For 32 GB VRAM use the 132.76 GB file. Memory estimates are approximate and include KV cache and runtime buffers.

What GPU do you need for DeepSeek V4 Flash 0731 IK?

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: 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 0731 IK quant should you choose?

If memory is tight, the smallest file is the first one in the table at 65.4 GB, with the quality trade-offs that come with fewer bits.

New to quant names? Read GGUF quantization explained.

Specifications

Architecture from the GGUF header

Architecturedeepseek4
Parameters284B

Source

Quantized byKeinNiemand
FamilyDeepSeek
Licensemit
Downloads (HF)8,091
Last updated2026-09-14

Frequently asked questions about DeepSeek V4 Flash 0731 IK

How much VRAM does DeepSeek V4 Flash 0731 IK need?

About 119.6 GB at its main file with an 8K context window. That is the 106.3 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 IK on an 8 GB GPU?

Not at an 8K context. The smallest file needs about 73.8 GB, so you would need a larger GPU or CPU offloading.

What is the file size of DeepSeek V4 Flash 0731 IK GGUF?

It depends on the quant: file is 65.4 GB, file is 73.2 GB, file is 93.0 GB, file is 93.4 GB, file is 106.3 GB, file is 122.0 GB and 3 more.

Where can I download DeepSeek V4 Flash 0731 IK GGUF?

The files are on Hugging Face in the KeinNiemand/DeepSeek-V4-Flash-0731-IK_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 IK 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 0731 IK on 8 GB VRAM?

The MXFP4 quantized file is 161.87 GB; it requires more than 8 GB VRAM. A smaller quantization is needed for 8 GB VRAM.

Does the model support vision tasks?

No, the model does not include vision capabilities.

What backend is required to load the GGUF files?

The files require a recent build of ik_llama.cpp that supports IK-specific IQK and IQT tensor types.

Other GGUFs of DeepSeek-V4-Flash-0731

Similar models

Browse all →

More from KeinNiemand

All by KeinNiemand →

More DeepSeek models

All DeepSeek →

Other large models

All large models →