DeepSeek V4 Flash 0731 REAP needs about 122.6 GB of memory at Q4_K_M (a 109.0 GB file) with an 8K context window, which is more than a 48 GB GPU holds. This repository has 3 quants, from 65.4 GB (Q2_K) to 109.0 GB (Q4_K_M).
DeepSeek V4 Flash 0731 REAP – a 193B parameter REAP‑pruned model in GGUF format
DeepSeek V4 Flash 0731 REAP is a 193B parameter model with 160 routed experts per layer. It is released under the MIT license and uses the deepseek4 architecture. Runtime support is experimental and may vary by GPU.
Good at
- REAP‑pruned expert efficiency
- Large context support up to 1,048,576 tokens
- MIT licensed for open use
Limits
- Experimental runtime compatibility
- Not recommended for Q2_K or Q4_K_M in llama.cpp
- GPU offload may fail on NVIDIA GB10
Best for
- Research on expert‑pruned models
- Testing new quantization schemes
- Exploring large‑context generation
Who it’s for. Developers and researchers interested in deepseek4 architecture and REAP‑pruned models. Suitable for those who can experiment with experimental runtimes and have access to compatible GPU hardware.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| Q2_K | 65.4 GB | 73.8 GB | Too big | Get file → | |
| MXFP4 | 101.2 GB | 113.8 GB | Too big | Get file → | |
| Q4_K_M | 109.0 GB | 122.6 GB | Too big | Get file → |
For 8 GB VRAM, use Q2_K (file size 65.43 GB, estimated 73.8 GB memory at 8 k context). For 12 GB VRAM, use MXFP4 (101.16 GB, 113.8 GB memory). For 16 GB VRAM, use Q4_K_M (109 GB, 122.6 GB memory). Estimates are approximate.
What GPU do you need for DeepSeek V4 Flash 0731 REAP?
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 REAP quant should you choose?
Q4_K_M is the usual starting point for DeepSeek V4 Flash 0731 REAP: it balances file size and quality well. Here it is a 109.0 GB file.
If memory is tight, the smallest file is Q2_K at 65.4 GB, with the quality trade-offs that come with fewer bits. Use it only when memory is very tight. Expect a clear drop in quality, which is usually worse on small models and easier to live with on large ones.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Frequently asked questions about DeepSeek V4 Flash 0731 REAP
How much VRAM does DeepSeek V4 Flash 0731 REAP need?
About 122.6 GB at Q4_K_M with an 8K context window. That is the 109.0 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 REAP 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 REAP GGUF?
It depends on the quant: Q2_K is 65.4 GB, MXFP4 is 101.2 GB, Q4_K_M is 109.0 GB.
What context length does DeepSeek V4 Flash 0731 REAP support?
The GGUF header lists a trained context length of 1,048,576 tokens (1024K).
Where can I download DeepSeek V4 Flash 0731 REAP GGUF?
The files are on Hugging Face in the heath0xFF/DeepSeek-V4-Flash-0731-REAP-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 REAP 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 Q2_K quantization requires about 73.8 GB memory at 8 k context, so 8 GB VRAM is insufficient.
Is this model suitable for production use?
Runtime support is experimental; it is not recommended for production without further testing.
What license does this model use?
The model is released under the MIT license.