DeepSeek R1 needs about 445.9 GB of memory at Q4_K_M (a 404.4 GB file) with an 8K context window, which is more than a 48 GB GPU holds. This repository has 24 quants, from 133.6 GB (IQ1_S) to 713.3 GB (Q8_0). Longer context needs more: at 32K tokens the KV cache alone adds about 163.7 GB.
DeepSeek R1 – a 671‑billion‑parameter reasoning model
DeepSeek R1 is a 671‑billion‑parameter model trained with large‑scale reinforcement learning. It focuses on reasoning tasks and is available in multiple GGUF quantizations. The model is released under the MIT license.
Good at
- Reasoning
- Chain‑of‑thought generation
- Self‑verification
Limits
- No vision support
- Requires large VRAM for higher quants
- No language metadata provided
Best for
- Math problem solving
- Code reasoning
- Long‑form logical explanations
Who it’s for. Researchers and developers who need a reasoning‑focused LLM and are comfortable managing large model files. The model is open‑source and can be run locally with llama.cpp.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| IQ1_S | 133.6 GB (4 parts) | 175.0 GB | Too big | Get file → | |
| IQ1_M | 148.9 GB (4 parts) | 190.3 GB | Too big | Get file → | |
| IQ2_XXS | 174.4 GB (5 parts) | 215.9 GB | Too big | Get file → | |
| IQ2_XS | 195.1 GB (5 parts) | 236.5 GB | Too big | Get file → | |
| IQ2_S | 197.0 GB (6 parts) | 238.4 GB | Too big | Get file → | |
| IQ2_M | 217.4 GB (6 parts) | 258.9 GB | Too big | Get file → | |
| Q2_K | 244.0 GB (7 parts) | 285.5 GB | Too big | Get file → | |
| Q2_K_L | 244.9 GB (7 parts) | 286.4 GB | Too big | Get file → | |
| IQ3_XXS | 257.9 GB (7 parts) | 299.4 GB | Too big | Get file → | |
| Q3_K_S | 289.1 GB (8 parts) | 330.5 GB | Too big | Get file → | |
| IQ3_M | 292.1 GB (8 parts) | 333.5 GB | Too big | Get file → | |
| Q3_K_M | 319.2 GB (9 parts) | 360.6 GB | Too big | Get file → | |
| Q3_K_L | 347.4 GB (9 parts) | 388.9 GB | Too big | Get file → | |
| Q3_K_XL | 348.3 GB (9 parts) | 389.7 GB | Too big | Get file → | |
| IQ4_XS | 357.1 GB (10 parts) | 398.6 GB | Too big | Get file → | |
| IQ4_NL | 378.1 GB (10 parts) | 419.5 GB | Too big | Get file → | |
| Q4_0 | 379.0 GB (10 parts) | 420.5 GB | Too big | Get file → | |
| Q4_K_S | 380.0 GB (10 parts) | 421.4 GB | Too big | Get file → | |
| Q4_K_M | 404.4 GB (11 parts) | 445.9 GB | Too big | Get file → | |
| Q4_1 | 419.9 GB (11 parts) | 461.4 GB | Too big | Get file → | |
| Q5_K_S | 461.8 GB (12 parts) | 503.3 GB | Too big | Get file → | |
| Q5_K_M | 475.4 GB (13 parts) | 516.8 GB | Too big | Get file → | |
| Q6_K | 550.8 GB (15 parts) | 592.2 GB | Too big | Get file → | |
| Q8_0 | 713.3 GB (20 parts) | 754.7 GB | Too big | Get file → |
For 8 GB VRAM, use IQ1_S or IQ1_M. For 12 GB VRAM, IQ2_XXS or IQ2_XS are suitable. For 16 GB VRAM, IQ2_S or IQ2_M provide a balance between size and performance. Larger quants such as Q4_K_M or Q5_K_S require 24 GB or more.
What GPU do you need for DeepSeek R1?
These figures add the file size, an fp16 KV cache at an 8K context window (worked out from this model’s own header) 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 R1 quant should you choose?
Q4_K_M is the usual starting point for DeepSeek R1: it balances file size and quality well. Here it is a 404.4 GB file.
The Q8_0 file is 713.3 GB, about 1.8 times the size of Q4_K_M. It keeps the weights closer to the original, which only matters if you have the memory to spare.
If memory is tight, the smallest file is IQ1_S at 133.6 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
Source
Frequently asked questions about DeepSeek R1
How much VRAM does DeepSeek R1 need?
About 445.9 GB at Q4_K_M with an 8K context window. That is the 404.4 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 R1 on an 8 GB GPU?
Not at an 8K context. The smallest file needs about 175.0 GB, so you would need a larger GPU or CPU offloading.
What is the file size of DeepSeek R1 GGUF?
It depends on the quant: IQ1_S is 133.6 GB, IQ1_M is 148.9 GB, IQ2_XXS is 174.4 GB, IQ2_XS is 195.1 GB, IQ2_S is 197.0 GB, IQ2_M is 217.4 GB and 18 more.
What context length does DeepSeek R1 support?
The GGUF header lists a trained context length of 163,840 tokens (160K). The KV cache grows with the context you use: about 40.9 GB at 8K and 163.7 GB at 32K.
Where can I download DeepSeek R1 GGUF?
The files are on Hugging Face in the bartowski/DeepSeek-R1-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 R1 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 R1 on 8 GB VRAM?
Use the IQ1_S or IQ1_M quant, which estimate 175–190.3 GB of memory at 8 k context. The file sizes are 133.56 GB and 148.88 GB respectively.
Is DeepSeek R1 licensed for commercial use?
The model is released under the MIT license, which allows commercial use.
Does DeepSeek R1 support vision tasks?
No, the model does not include vision capabilities.