DeepSeek V4 Flash MTP needs about 160.4 GB of memory at Q3_K_M (a 142.8 GB file) with an 8K context window, which is more than a 48 GB GPU holds. This repository has 4 quants, from 142.8 GB (Q3_K_M) to 160.0 GB.
DeepSeek V4 Flash MTP – 291B parameter LLM with million‑token context
DeepSeek V4 Flash MTP is a 291B parameter language model built on the deepseek4 architecture. It supports a one‑million‑token context length and uses a draft‑head multi‑token‑prediction mechanism for efficient decoding. The model is released under the MIT license.
Good at
- Large‑context reasoning
- High‑token‑rate decoding
- Versatile domain knowledge
Limits
- Requires 160.4 GB VRAM at 8 k context (approx.)
- No vision or audio capabilities
- No official quantization beyond Q3_K_M and unknown
Best for
- Long‑form content generation
- Complex reasoning tasks
- Research prototyping
Who it’s for. Developers and researchers needing a large‑context open‑source model with efficient decoding. The model is suitable for GPU‑heavy workloads and can be used with llama.cpp on stock mainline after 2026‑08‑13 merge.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| Q3_K_M | 142.8 GB (4 parts) | 160.4 GB | Too big | Get file → | |
| Q3_K_M | 142.8 GB (4 parts) | 160.4 GB | Too big | Get file → | |
| — | 160.0 GB (5 parts) | 179.7 GB | Too big | Get file → | |
| — | 160.0 GB (4 parts) | 179.7 GB | Too big | Get file → |
For 8 GB VRAM, the Q3_K_M quant (142.77 GB file) is too large; the model is not feasible on such hardware. For 16 GB VRAM, the Q3_K_M quant remains too large. The only practical option is to use a higher‑quantization level not listed here or run on a multi‑GPU setup with at least 160.4 GB VRAM. The estimates are approximate.
What GPU do you need for DeepSeek V4 Flash MTP?
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 MTP quant should you choose?
Q3_K_M (142.8 GB) is the middle option in this repository. A reasonable choice when a 4-bit file will not fit, mostly on larger models. Quality loss is noticeable but often acceptable.
If memory is tight, the smallest file is Q3_K_M at 142.8 GB, with the quality trade-offs that come with fewer bits. A reasonable choice when a 4-bit file will not fit, mostly on larger models. Quality loss is noticeable but often acceptable.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Source
Frequently asked questions about DeepSeek V4 Flash MTP
How much VRAM does DeepSeek V4 Flash MTP need?
About 160.4 GB at Q3_K_M with an 8K context window. That is the 142.8 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 MTP on an 8 GB GPU?
Not at an 8K context. The smallest file needs about 160.4 GB, so you would need a larger GPU or CPU offloading.
What is the file size of DeepSeek V4 Flash MTP GGUF?
It depends on the quant: Q3_K_M is 142.8 GB, Q3_K_M is 142.8 GB, file is 160.0 GB, file is 160.0 GB.
Where can I download DeepSeek V4 Flash MTP GGUF?
The files are on Hugging Face in the rogerai-fyi/DeepSeek-V4-Flash-MTP-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 MTP 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 MTP on 8 GB VRAM?
No. The model requires approximately 160.4 GB VRAM at 8 k context, far exceeding 8 GB.
Does the model support vision or audio input?
No. The model has no vision or audio capabilities.
What license does the model use?
The model is released under the MIT license.