How much memory do I need to run DeepSeek-R1 (Q4_K_M)?

The file size is 404 GB (measured). With a 8,192-token context, the estimated memory including the KV cache is 446 GB. The formula is published at https://localllm.az-arte.com/en/guide/memory-calculation.

Does DeepSeek-R1 run in 8 GB, 12 GB, 16 GB or 24 GB of VRAM?

8 GB: no quantization fits; 12 GB: no quantization fits; 16 GB: no quantization fits; 24 GB: no quantization fits (estimated at 8,192 tokens with an f16 KV cache, files from the most-downloaded source).

What is the command to run DeepSeek-R1 with Ollama?

Run "ollama run hf.co/unsloth/DeepSeek-R1-GGUF:Q4_K_M" to use the GGUF on Hugging Face directly. To use another quantization, replace the name at the end.

deepseek-ai

DeepSeek-R1

Parameters: 671B
Max context length: 163,840
Architecture: deepseek2
License: Commercial OK

Base model: deepseek-ai/DeepSeek-R1
License: mit

Memory needed to run Q4_K_M at 8K tokens

446 GB Estimated

File size 404 GB Measured

VRAM it fits in (rough guide)

8 GB does not fit
12 GB does not fit
16 GB does not fit
24 GB does not fit
32 GB does not fit
48 GB does not fit

The smallest, UD-IQ1_S, needs an estimated 182 GB at 8K.

Macs cap how much memory the GPU can use, so leave headroom in unified memory.

Data as of Nov 11, 2026

About DeepSeek-R1

DeepSeek-R1 is a GGUF model based on deepseek-ai/DeepSeek-R1 on Hugging Face. GGUF files are published by unsloth/DeepSeek-R1-GGUF. It has about 671B parameters, the deepseek2 architecture, a maximum context length of 163,840 tokens (values written in the GGUF).

There are 12 quantizations, with file sizes from 140 GB (UD-IQ1_S) to 1,342 GB (BF16) (actual sizes on Hugging Face).

Running Q4_K_M (404 GB) with an 8,192-token context needs an estimated 446 GB of memory, and an estimated 569 GB at 32,768 tokens. The smallest, UD-IQ1_S (140 GB), needs an estimated 182 GB at 8,192 tokens.

The original model is licensed under MIT License, which allows commercial use (keep the license text and copyright notice when redistributing). Check the original license for the terms of use.

File size and estimated memory per quantization

Up to 8 GB
Up to 16 GB
Up to 24 GB
Up to 48 GB
Up to 128 GB
More

File sizes are the actual values on Hugging Face. Memory is estimated for each context length (KV cache f16). The formula is in How memory is calculated.

All memory figures are Estimated (KV cache f16). It should run on a setup with at least this much VRAM (or Mac unified memory). Click a value to see the devices and models that fit that amount (Find by GPU / Mac).

About the "quality guide" (how far a quantization drifts from the original)

For each quantization type, how far the output drifts from the original model, grouped into six levels from values llama.cpp measured on Llama-3-8B.

The value is mean KLD (the difference between the original and quantized model in next-token probabilities); smaller means closer to the original. Quantizations made with an imatrix (importance data) come out slightly smaller. Levels use the value without an imatrix (or with one, for types that only have that).

These values were not measured on this model. Larger models and MoE models tend to drift less with the same type. Distributor-specific mixed types (such as UD-Q4_K_XL) cannot be matched by type name alone, so they show "-".

Quantization	Quality guide	KLD (no imatrix)	KLD (imatrix)	Size (Llama-3-8B)
Q8_0	Nearly identical	0.0014	-	7.96 GiB
Q6_K	Very small	0.0055	-	6.14 GiB
Q5_K_M	Very small	0.0108	-	5.33 GiB
Q5_K_S	Small	0.0166	-	5.21 GiB
Q5_1	Small	0.0180	-	5.65 GiB
Q5_0	Small	0.0222	-	5.21 GiB
Q4_K_M	Small	0.0313	0.0282	4.58 GiB
Q4_K_S	Medium	0.0431	0.0320	4.37 GiB
IQ4_NL	Small	-	0.0357	4.35 GiB
IQ4_XS	Small	-	0.0363	4.14 GiB
Q4_1	Medium	0.0717	-	4.78 GiB
Q4_0	Medium	0.0719	-	4.34 GiB
Q3_K_L	Medium	0.0872	0.0731	4.03 GiB
Q3_K_M	Medium	0.1019	0.0844	3.74 GiB
IQ3_M	Medium	-	0.1025	3.53 GiB
IQ3_S	Medium	-	0.1113	3.42 GiB
IQ3_XS	Large	-	0.1387	3.28 GiB
IQ3_XXS	Large	-	0.1836	3.05 GiB
Q3_K_S	Large	0.2282	0.1998	3.41 GiB
IQ2_M	Large	-	0.3260	2.74 GiB
Q2_K	Very large	0.4451	0.3322	2.96 GiB
Q2_K_S	Very large	-	0.4034	2.96 GiB
IQ2_S	Very large	-	0.4392	2.56 GiB
IQ2_XS	Very large	-	0.5463	2.43 GiB
IQ2_XXS	Very large	-	0.8120	2.24 GiB
IQ1_M	Very large	-	1.3931	2.01 GiB
IQ1_S	Very large	-	2.2546	1.88 GiB

Source: llama.cpp (MIT), measured on Llama-3-8B against FP16, Wikitext-2, revision f364eb6f. Original table · Reading and choosing quantizations

GGUF source

unsloth/DeepSeek-R1-GGUF

Converted and published by a third party About this label

Last checked: Nov 11, 2026

Architecture: deepseek2; Supported before llama.cpp b4409
Parameters: 671.0B
Max context length: 163,840
License: mit Commercial OK

File size and memory per quantization for DeepSeek-R1 (unsloth/DeepSeek-R1-GGUF)
Quantization	File size Measured	Quality guide	4,096	8,192	16,384	32,768	65,536	131,072	163,840 max
UD-IQ1_S (3 parts)	140 GB	-	162 GB	182 GB	223 GB	305 GB	469 GB	796 GB	960 GB
UD-IQ1_M (4 parts)	169 GB	-	190 GB	211 GB	252 GB	333 GB	497 GB	825 GB	988 GB
UD-IQ2_XXS (4 parts)	196 GB	-	217 GB	238 GB	279 GB	361 GB	524 GB	852 GB	1,016 GB
Q2_K (5 parts) DeepSeek-R1-Q2_K_XS	221 GB	Very large	243 GB	263 GB	304 GB	386 GB	550 GB	877 GB	1,041 GB
UD-Q2_K_XL (5 parts)	227 GB	-	248 GB	268 GB	309 GB	391 GB	555 GB	882 GB	1,046 GB
Q2_K (5 parts) DeepSeek-R1-Q2_K	244 GB	Very large	265 GB	286 GB	327 GB	409 GB	572 GB	900 GB	1,064 GB
Q2_K_L (5 parts)	244 GB	-	266 GB	286 GB	327 GB	409 GB	573 GB	900 GB	1,064 GB
Q3_K_M (7 parts)	319 GB	Medium	340 GB	361 GB	402 GB	484 GB	647 GB	975 GB	1,139 GB
Q4_K_M (9 parts)	404 GB	Small	426 GB	446 GB	487 GB	569 GB	733 GB	1,060 GB	1,224 GB
Q5_K_M (10 parts)	475 GB	Very small	497 GB	517 GB	558 GB	640 GB	804 GB	1,131 GB	1,295 GB
Q6_K (12 parts)	551 GB	Very small	572 GB	593 GB	633 GB	715 GB	879 GB	1,207 GB	1,370 GB
Q8_0 (15 parts)	713 GB	Nearly identical	735 GB	755 GB	796 GB	878 GB	1,042 GB	1,369 GB	1,533 GB
BF16 (30 parts)	1,342 GB	Original precision	1,364 GB	1,384 GB	1,425 GB	1,507 GB	1,671 GB	1,998 GB	2,162 GB

Ollama  ollama run hf.co/unsloth/DeepSeek-R1-GGUF:Q4_K_M

llama.cpp  llama-server -hf unsloth/DeepSeek-R1-GGUF:Q4_K_M

Third-party evaluations

Values from other organizations for the original model before quantization. Not this site's own rating. Values are for the original model before quantization. GGUF quantization (especially small quantizations) can lower performance.

Human preference (LMArena)

1,398 1,393–1,403

deepseek-r1 · 18,524 votes · Values as of Nov 9, 2026

Source: LMArena (CC BY 4.0). Original leaderboard · See in ranking

Capabilities (Epoch AI)

138.7 136.7–140.4

DeepSeek-R1 · Values as of Jan 20, 2025

Source: Epoch Capabilities Index (Epoch AI) (CC BY 4.0). Original leaderboard · See in ranking

Hard science (GPQA)

69.2% 63.2%–75.2%

DeepSeek-R1 · Values as of May 26, 2025

Source: GPQA Diamond (Epoch AI) (CC BY 4.0). Original leaderboard · See in ranking

License and commercial use

MIT License Commercial OK

Allows commercial use, modification and redistribution. When redistributing, keep the copyright notice and license text.

Based on the license on the original model's Hugging Face page. Original license text (checked Oct 18, 2026)

The summary states facts this site read from the original text and is not legal advice. Always check the original license and its latest terms before use. All licenses

How to run

Example commands using Q4_K_M. To use another quantization, replace the name with one from the table above. How to run local LLMs

Ollama

Run it by pointing directly at the GGUF on Hugging Face.

ollama run hf.co/unsloth/DeepSeek-R1-GGUF:Q4_K_M

llama.cpp

Pass the repository and quantization name to llama-server (or llama-cli) with -hf.

llama-server -hf unsloth/DeepSeek-R1-GGUF:Q4_K_M

LM Studio

Paste unsloth/DeepSeek-R1-GGUF into the search box, download it, and choose the quantization from the file list.

Frequently asked questions

How much memory do I need to run DeepSeek-R1 (Q4_K_M)?: The file size is 404 GB (measured). With a 8,192-token context, the estimated memory including the KV cache is 446 GB. The formula is published at https://localllm.az-arte.com/en/guide/memory-calculation.
Can DeepSeek-R1 be used commercially?: The original model's license is MIT License (Commercial use allowed).Allows commercial use, modification and redistribution. When redistributing, keep the copyright notice and license text. Check the original license text before deciding.
Does DeepSeek-R1 run in 8 GB, 12 GB, 16 GB or 24 GB of VRAM?: 8 GB: no quantization fits; 12 GB: no quantization fits; 16 GB: no quantization fits; 24 GB: no quantization fits (estimated at 8,192 tokens with an f16 KV cache, files from the most-downloaded source).
What is the command to run DeepSeek-R1 with Ollama?: Run "ollama run hf.co/unsloth/DeepSeek-R1-GGUF:Q4_K_M" to use the GGUF on Hugging Face directly. To use another quantization, replace the name at the end.