Skip to content
localllm
日本語

openai

gpt-oss-120b

Parameters
116.8B
Max context length
131,072
Architecture
gpt-oss
License
Commercial OK
Base model
openai/gpt-oss-120b
License
apache-2.0
llama.cpp
b6096+

About gpt-oss-120b

gpt-oss-120b is a GGUF model based on openai/gpt-oss-120b on Hugging Face. GGUF files are published by 3 sources (ggml-org/gpt-oss-120b-GGUF, unsloth/gpt-oss-120b-GGUF, lmstudio-community/gpt-oss-120b-GGUF). It has about 116.8B parameters, the gpt-oss architecture, a maximum context length of 131,072 tokens (values written in the GGUF).

There are 17 quantizations, with file sizes from 62.6 GB (Q3_K_S) to 65.4 GB (F16) (actual sizes on Hugging Face).

Running MXFP4 (63.4 GB) with an 8,192-token context needs an estimated 64.7 GB of memory, and an estimated 65.6 GB at 32,768 tokens. For 8,192 tokens, a GPU with 64 GB or more of VRAM, or a Mac with 64 GB or more of unified memory, is a rough guide (macOS caps how much memory the GPU can use, so leave some headroom). The smallest, Q3_K_S (62.6 GB), needs an estimated 63.8 GB at 8,192 tokens.

The original model is licensed under Apache License 2.0, 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).

3 sources publish GGUF files of this model. Quantization types and file sizes differ slightly between sources, so there is a table for each source below.

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

ggml-org/gpt-oss-120b-GGUF

Converted by the llama.cpp developers About this label

Last checked: Nov 11, 2026

Architecture
gpt-oss
llama.cpp b6096 and later
Parameters
116.8B
Max context length
131,072
License
apache-2.0 Commercial OK
File size and memory per quantization for gpt-oss-120b (ggml-org/gpt-oss-120b-GGUF)
Quantization File size Measured Quality guide 4,096 8,192 16,384 32,768 65,536 131,072 max
MXFP4 63.4 GB - 64.5 GB 64.7 GB 65.0 GB 65.6 GB 66.8 GB 69.2 GB 
Ollama  ollama run hf.co/ggml-org/gpt-oss-120b-GGUF:MXFP4
llama.cpp  llama-server -hf ggml-org/gpt-oss-120b-GGUF:MXFP4

GGUF source

unsloth/gpt-oss-120b-GGUF

Converted and published by a third party About this label

Last checked: Nov 11, 2026

Architecture
gpt-oss
llama.cpp b6096 and later
Parameters
116.8B
Max context length
131,072
License
apache-2.0 Commercial OK
File size and memory per quantization for gpt-oss-120b (unsloth/gpt-oss-120b-GGUF)
Quantization File size Measured Quality guide 4,096 8,192 16,384 32,768 65,536 131,072 max
Q3_K_S (2 parts) 62.6 GB Large 63.7 GB 63.8 GB 64.1 GB 64.7 GB 66.0 GB 68.4 GB
Q2_K (2 parts) 62.6 GB Very large 63.7 GB 63.8 GB 64.2 GB 64.8 GB 66.0 GB 68.4 GB
Q4_0 (2 parts) 62.6 GB Medium 63.7 GB 63.9 GB 64.2 GB 64.8 GB 66.0 GB 68.4 GB
Q3_K_M (2 parts) 62.6 GB Medium 63.8 GB 63.9 GB 64.2 GB 64.8 GB 66.0 GB 68.4 GB
Q4_1 (2 parts) 62.7 GB Medium 63.8 GB 64.0 GB 64.3 GB 64.9 GB 66.1 GB 68.5 GB
Q4_K_S (2 parts) 62.8 GB Medium 63.9 GB 64.0 GB 64.3 GB 64.9 GB 66.2 GB 68.6 GB
Q4_K_M (2 parts) 62.8 GB Small 63.9 GB 64.0 GB 64.3 GB 65.0 GB 66.2 GB 68.6 GB
Q2_K_L (2 parts) 62.9 GB - 64.0 GB 64.1 GB 64.4 GB 65.0 GB 66.3 GB 68.7 GB
Q5_K_S (2 parts) 62.9 GB Small 64.0 GB 64.2 GB 64.5 GB 65.1 GB 66.3 GB 68.7 GB
Q5_K_M (2 parts) 62.9 GB Very small 64.0 GB 64.2 GB 64.5 GB 65.1 GB 66.3 GB 68.7 GB
UD-Q4_K_XL (2 parts) 63.0 GB - 64.1 GB 64.3 GB 64.6 GB 65.2 GB 66.4 GB 68.8 GB
Q6_K (2 parts) 63.3 GB Very small 64.4 GB 64.6 GB 64.9 GB 65.5 GB 66.7 GB 69.1 GB
UD-Q6_K_XL (2 parts) 63.3 GB - 64.4 GB 64.6 GB 64.9 GB 65.5 GB 66.7 GB 69.1 GB
Q8_0 (2 parts) 63.4 GB Nearly identical 64.5 GB 64.7 GB 65.0 GB 65.6 GB 66.8 GB 69.2 GB
UD-Q8_K_XL (2 parts) 64.5 GB - 65.6 GB 65.8 GB 66.1 GB 66.7 GB 67.9 GB 70.3 GB
F16 65.4 GB Original precision 66.5 GB 66.6 GB 67.0 GB 67.6 GB 68.8 GB 71.2 GB 
Ollama  ollama run hf.co/unsloth/gpt-oss-120b-GGUF:Q4_K_M
llama.cpp  llama-server -hf unsloth/gpt-oss-120b-GGUF:Q4_K_M

GGUF source

lmstudio-community/gpt-oss-120b-GGUF

Converted and published by a third party About this label

Last checked: Nov 11, 2026

Architecture
gpt-oss
llama.cpp b6096 and later
Parameters
116.8B
Max context length
131,072
License
apache-2.0 Commercial OK
File size and memory per quantization for gpt-oss-120b (lmstudio-community/gpt-oss-120b-GGUF)
Quantization File size Measured Quality guide 4,096 8,192 16,384 32,768 65,536 131,072 max
MXFP4 (2 parts) 63.4 GB - 64.5 GB 64.7 GB 65.0 GB 65.6 GB 66.8 GB 69.2 GB 
Ollama  ollama run hf.co/lmstudio-community/gpt-oss-120b-GGUF:MXFP4
llama.cpp  llama-server -hf lmstudio-community/gpt-oss-120b-GGUF:MXFP4

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)

1,352 1,347–1,356

gpt-oss-120b · 30,133 votes · Values as of Nov 9, 2026

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

Capabilities (Epoch AI)

140.2 138.3–142.0

gpt-oss-120b · Values as of Aug 5, 2025

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

License and commercial use

Apache License 2.0 Commercial OK

Allows commercial use, modification and redistribution. When redistributing, keep the license text and copyright notice, and note changes in modified files.

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

Mainline llama.cpp can load this model's format (gpt-oss) from b6096 (Aug 5, 2025) onward. Ollama and LM Studio bundle llama.cpp internally, so an older app fails to load it with errors such as "unknown model architecture". If that happens, update the app.

Example commands using MXFP4. To use another quantization, replace the name with one from the table above. Commands for files from other sources are in each source's section. How to run local LLMs

Ollama

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

ollama run hf.co/ggml-org/gpt-oss-120b-GGUF:MXFP4

llama.cpp

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

llama-server -hf ggml-org/gpt-oss-120b-GGUF:MXFP4

LM Studio

Paste ggml-org/gpt-oss-120b-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 gpt-oss-120b (MXFP4)?
The file size is 63.4 GB (measured). With a 8,192-token context, the estimated memory including the KV cache is 64.7 GB. The formula is published at https://localllm.az-arte.com/en/guide/memory-calculation.
Can gpt-oss-120b be used commercially?
The original model's license is Apache License 2.0 (Commercial use allowed).Allows commercial use, modification and redistribution. When redistributing, keep the license text and copyright notice, and note changes in modified files. Check the original license text before deciding.
Does gpt-oss-120b 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 gpt-oss-120b with Ollama?
Run "ollama run hf.co/ggml-org/gpt-oss-120b-GGUF:MXFP4" to use the GGUF on Hugging Face directly. To use another quantization, replace the name at the end.