DeepSeek V4 Flash 0731 Abliterated DS4 Headroom128 needs about 7.2 GB of memory at its main file (a 6.0 GB file) with an 8K context window, so it fits on a 8 GB GPU. This repository has 2 quants, from 6.0 GB to 86.7 GB.
DeepSeek V4 Flash 0731 Abliterated DS4 Headroom128 – a large‑context, DS4‑ready GGUF model
This model is a rank‑1 abliterated derivative of DeepSeek‑V4‑Flash‑0731. It is released as a mixed‑precision GGUF conversion with DS4 support. The artifact contains two GGUF files totaling about 86.34 GiB and retains the 1,048,576‑token training context declaration.
Good at
- Large one‑million‑token context
- DS4 three‑stage DSpark support
- DeepSeek‑4 architecture
Limits
- Requires >7.2 GB VRAM for the 5.99 GB GGUF
- Large memory footprint for 128 GB host
- No vision or multimodal support
Best for
- Agents needing long‑context reasoning
- Research on DS4‑enabled models
- Deployments on 128 GB Apple Silicon
Who it’s for. Engineers and researchers who need a DeepSeek‑V4‑Flash model with DS4 support and a large context window. The model is experimental and not yet benchmarked for generation quality or throughput.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| — | 6.0 GB | 7.2 GB | Tight | Get file → | |
| — | 86.7 GB | 97.6 GB | Too big | Get file → |
For systems with limited VRAM, use the 5.99 GB GGUF file; it requires about 7.2 GB VRAM at 8 k context. For full DS4 runtime and KV cache on a 128 GB host, use the 86.72 GB GGUF file, which estimates 97.6 GB VRAM at 8 k context. Memory estimates are approximate.
What GPU do you need for DeepSeek V4 Flash 0731 Abliterated DS4 Headroom128?
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: the best quant that fits is the main file (6.0 GB file, about 7.2 GB in use).
- 12 GB VRAM: the best quant that fits is the main file (6.0 GB file, about 7.2 GB in use).
- 16 GB VRAM: the best quant that fits is the main file (6.0 GB file, about 7.2 GB in use).
- 24 GB VRAM: the best quant that fits is the main file (6.0 GB file, about 7.2 GB in use).
- 32 GB VRAM: the best quant that fits is the main file (6.0 GB file, about 7.2 GB in use).
- 48 GB VRAM: the best quant that fits is the main file (6.0 GB file, about 7.2 GB in use).
Specifications
Architecture from the GGUF header
Frequently asked questions about DeepSeek V4 Flash 0731 Abliterated DS4 Headroom128
How much VRAM does DeepSeek V4 Flash 0731 Abliterated DS4 Headroom128 need?
About 7.2 GB at its main file with an 8K context window. That is the 6.0 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 0731 Abliterated DS4 Headroom128 on an 8 GB GPU?
Yes, with the smaller files: the 6.0 GB file needs about 7.2 GB at 8K context. Larger quants in this repository do not fit in 8 GB.
What is the file size of DeepSeek V4 Flash 0731 Abliterated DS4 Headroom128 GGUF?
It depends on the quant: file is 6.0 GB, file is 86.7 GB.
Where can I download DeepSeek V4 Flash 0731 Abliterated DS4 Headroom128 GGUF?
The files are on Hugging Face in the apetersson/DeepSeek-V4-Flash-0731-Abliterated-DS4-Headroom128 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 Abliterated DS4 Headroom128 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 5.99 GB GGUF requires about 7.2 GB VRAM at 8 k context, so it can run on 8 GB VRAM with some headroom. The 86.72 GB GGUF would not fit.
Does this model support vision or multimodal inputs?
No, the model has no vision or multimodal support.
What context length does it support?
It retains the 1,048,576‑token training context declaration, but inference has not been exercised.