DE

DeepSeek V4 Flash DSpark drafter

DeepSeekGGUF by bleysgUpdated 2026-09-03
Download on Hugging Face
Parameters19.8B
Context—
Layers—
Architecturedeepseek4-dspark
Smallest quant6.9 GB
Quick answer

DeepSeek V4 Flash DSpark drafter needs about 8.3 GB of memory at its main file (a 6.9 GB file) with an 8K context window, so it fits on a 12 GB GPU. This repository has 4 quants, from 6.9 GB to 7.0 GB.

DeepSeek V4 Flash DSpark drafter – a lightweight draft‑generation model

The DeepSeek V4 Flash DSpark drafter is a 19.8 billion‑parameter model designed for lossless speculative decoding. It contains three transformer layers and a 256‑expert MoE, optimized for the ds4 inference engine.

Good at

  • Lossless speculative decoding
  • Fast draft generation
  • Efficient 1‑M token context support

Limits

  • Requires ~8.3 GB VRAM at 8k context
  • No vision or multimodal support
  • MIT license only

Best for

  • Real‑time text drafting
  • Low‑latency chat assistants
  • Batch speculative decoding

Who it’s for. Developers building lightweight inference pipelines who need fast, lossless draft generation. The model is suitable for integration with the ds4 C+CUDA engine and for users who prefer open‑source solutions.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
—6.9 GB 8.3 GB
Too big Get file →
—6.9 GB 8.3 GB
Too big Get file →
—7.0 GB 8.3 GB
Too big Get file →
—7.0 GB 8.3 GB
Too big Get file →

All available GGUF files are quantized to an unknown precision. The file size is 6.94 GB (or 6.97 GB) and the estimated VRAM usage is 8.3 GB at 8k context. These estimates are approximate. No alternative quantizations are listed.

What GPU do you need for DeepSeek V4 Flash DSpark drafter?

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: the best quant that fits is the main file (7.0 GB file, about 8.3 GB in use).

Which DeepSeek V4 Flash DSpark drafter quant should you choose?

If memory is tight, the smallest file is the first one in the table at 6.9 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-dspark
Parameters19.8B

Source

Quantized bybleysg
FamilyDeepSeek
Licensemit
Downloads (HF)761
Last updated2026-09-03

Frequently asked questions about DeepSeek V4 Flash DSpark drafter

How much VRAM does DeepSeek V4 Flash DSpark drafter need?

About 8.3 GB at its main file with an 8K context window. That is the 6.9 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 DSpark drafter on an 8 GB GPU?

Not at an 8K context. The smallest file needs about 8.3 GB, so you would need a larger GPU or CPU offloading.

What is the file size of DeepSeek V4 Flash DSpark drafter GGUF?

It depends on the quant: file is 6.9 GB, file is 6.9 GB, file is 7.0 GB, file is 7.0 GB.

Where can I download DeepSeek V4 Flash DSpark drafter GGUF?

The files are on Hugging Face in the bleysg/DeepSeek-V4-Flash-DSpark-drafter-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 DSpark drafter 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?

The model requires approximately 8.3 GB VRAM at 8k context, so 8 GB is insufficient.

Does this model support vision tasks?

No, the model has no vision or multimodal capabilities.

What license does it use?

The model is released under the MIT license.

Other GGUFs of DeepSeek-V4-Flash

Similar models

Browse all →

More DeepSeek models

All DeepSeek →

Other medium models

All medium models →