GPT-OSS · Over 80B · Mixture of experts
GPT OSS 120B
116.83 billion parameters in total, but only 5.7 billion are read for each token. That split is the whole point of the design: it costs the memory of a large model and the speed of a small one. 18 of its 36 layers use a sliding window of 128 tokens, so only 18 keep growing with the conversation.
About GPT OSS 120B
The larger of OpenAI's two open-weight models, Apache 2.0, with about 5.7B parameters active per token. The native MXFP4 file is 59 GiB, so it needs 64 GB of memory or an expert split between a GPU and fast system RAM.
GPT OSS 120B is a mixture-of-experts model with 116.83B parameters, 5.7B of them active for each token, built from 36 layers with a hidden size of 2880. Its repository was published on 2025-08-04. The license listed on Hugging Face is apache-2.0. It is released in a native MXFP4 format, so the download is 59.03 GiB however it is labeled. 18 of its 36 layers use a sliding window of 128 tokens, so only 18 keep growing with the conversation.
Memory needed, by quantization
At an 8k context window. Quality is our rough ranking of how much the compression costs you: anything at or above Q5 is hard to tell apart from the original in normal use.
| Quantization | File size | KV cache | Total VRAM | Quality | What it costs you |
|---|---|---|---|---|---|
| MXFP4 | 59.03 GB | 0.29 GB | 59.4 GB | 98% | The 4-bit format some models, gpt-oss among them, are released in. It is their native precision, so there is no quality step down. |
What a longer conversation costs
Same model at MXFP4, only the context window changes. 18 of its 36 layers use sliding-window attention, so their cache stops growing past 128 tokens and only the other 18 keep growing.
| Context | KV cache | Total VRAM | Fits in 8 GB | Fits in 12 GB | Fits in 24 GB |
|---|---|---|---|---|---|
| 4k | 0.15 GB | 59.26 GB | no | no | no |
| 8k | 0.29 GB | 59.4 GB | no | no | no |
| 16k | 0.57 GB | 59.68 GB | no | no | no |
| 32k | 1.13 GB | 60.24 GB | no | no | no |
| 64k | 2.25 GB | 61.36 GB | no | no | no |
| 128k | 4.5 GB | 63.61 GB | no | no | no |
The "fits" columns allow for the roughly 0.8 GB Windows keeps for the desktop.
Which hardware runs GPT OSS 120B
14 of 118 consumer devices run it at a quantization worth using, at an 8k context window.
| Device | Memory | Verdict | Quantization | Used | Speed |
|---|---|---|---|---|---|
| RTX PRO 6000 Blackwell NVIDIA | 96 GB | Runs great | MXFP4 | 59.4 GB | 242.6 t/s |
| Apple M3 Ultra 96GB Apple | 96 GB | Runs great | MXFP4 | 59.65 GB | 102.5 t/s |
| Apple M3 Ultra 256GB Apple | 256 GB | Runs great | MXFP4 | 59.65 GB | 102.5 t/s |
| Apple M3 Ultra 512GB Apple | 512 GB | Runs great | MXFP4 | 59.65 GB | 102.5 t/s |
| Apple M1 Ultra 128GB Apple | 128 GB | Runs great | MXFP4 | 59.65 GB | 100.1 t/s |
| Apple M2 Ultra 128GB Apple | 128 GB | Runs great | MXFP4 | 59.65 GB | 100.1 t/s |
| Apple M2 Ultra 192GB Apple | 192 GB | Runs great | MXFP4 | 59.65 GB | 100.1 t/s |
| Apple M4 Max 128GB Apple | 128 GB | Runs great | MXFP4 | 59.65 GB | 68.3 t/s |
| Apple M2 Max 96GB Apple | 96 GB | Runs great | MXFP4 | 59.65 GB | 50 t/s |
| Apple M3 Max 96GB Apple | 96 GB | Runs great | MXFP4 | 59.65 GB | 50 t/s |
| Apple M3 Max 128GB Apple | 128 GB | Runs great | MXFP4 | 59.65 GB | 50 t/s |
| NVIDIA DGX Spark 128GB NVIDIA | 128 GB | Runs great | MXFP4 | 59.68 GB | 37 t/s |
| Ryzen AI Max+ 395 96GB AMD | 96 GB | Runs great | MXFP4 | 59.68 GB | 26.8 t/s |
| Ryzen AI Max+ 395 128GB AMD | 128 GB | Runs great | MXFP4 | 59.68 GB | 26.8 t/s |
| RTX 6000 Ada Generation NVIDIA | 48 GB | Partial offload 36/36 | n/a | 59.4 GB | 53.1 t/s |
| RTX A6000 NVIDIA | 48 GB | Partial offload 36/36 | n/a | 59.4 GB | 49.1 t/s |
| Radeon PRO W7900 AMD | 48 GB | Partial offload 36/36 | n/a | 59.4 GB | 48.3 t/s |
| GeForce RTX 5090 NVIDIA | 32 GB | Partial offload 36/36 | n/a | 59.4 GB | 31.2 t/s |
| GeForce RTX 4090 NVIDIA | 24 GB | Partial offload 36/36 | n/a | 59.4 GB | 23.9 t/s |
| GeForce RTX 3090 Ti NVIDIA | 24 GB | Partial offload 36/36 | n/a | 59.4 GB | 23.9 t/s |
| GeForce RTX 3090 NVIDIA | 24 GB | Partial offload 36/36 | n/a | 59.4 GB | 23.7 t/s |
| GeForce RTX 5090 Laptop NVIDIA | 24 GB | Partial offload 36/36 | n/a | 59.4 GB | 23.6 t/s |
| Radeon RX 7900 XTX AMD | 24 GB | Partial offload 36/36 | n/a | 59.4 GB | 23.3 t/s |
| RTX A5000 NVIDIA | 24 GB | Partial offload 36/36 | n/a | 59.4 GB | 23.2 t/s |
| Radeon RX 7900 XT AMD | 20 GB | Partial offload 36/36 | n/a | 59.4 GB | 21 t/s |
Show 35 more devices
| GeForce RTX 5080 NVIDIA | 16 GB | Partial offload 36/36 | n/a | 59.4 GB | 20.1 t/s |
| GeForce RTX 5070 Ti NVIDIA | 16 GB | Partial offload 36/36 | n/a | 59.4 GB | 20 t/s |
| GeForce RTX 5080 Laptop NVIDIA | 16 GB | Partial offload 36/36 | n/a | 59.4 GB | 19.7 t/s |
| GeForce RTX 4080 SUPER NVIDIA | 16 GB | Partial offload 36/36 | n/a | 59.4 GB | 19.6 t/s |
| GeForce RTX 4080 NVIDIA | 16 GB | Partial offload 36/36 | n/a | 59.4 GB | 19.6 t/s |
| GeForce RTX 4070 Ti SUPER NVIDIA | 16 GB | Partial offload 36/36 | n/a | 59.4 GB | 19.5 t/s |
| GeForce RTX 4090 Laptop NVIDIA | 16 GB | Partial offload 36/36 | n/a | 59.4 GB | 19.1 t/s |
| Radeon RX 9070 XT AMD | 16 GB | Partial offload 36/36 | n/a | 59.4 GB | 19 t/s |
| Radeon RX 9070 AMD | 16 GB | Partial offload 36/36 | n/a | 59.4 GB | 19 t/s |
| Radeon RX 7800 XT AMD | 16 GB | Partial offload 36/36 | n/a | 59.4 GB | 18.9 t/s |
| Radeon RX 7900 GRE AMD | 16 GB | Partial offload 36/36 | n/a | 59.4 GB | 18.7 t/s |
| GeForce RTX 3080 Ti NVIDIA | 12 GB | Partial offload 36/36 | n/a | 59.4 GB | 18.6 t/s |
| GeForce RTX 3080 12GB NVIDIA | 12 GB | Partial offload 36/36 | n/a | 59.4 GB | 18.6 t/s |
| GeForce RTX 5060 Ti 16GB NVIDIA | 16 GB | Partial offload 36/36 | n/a | 59.4 GB | 18.5 t/s |
| RTX A4000 NVIDIA | 16 GB | Partial offload 36/36 | n/a | 59.4 GB | 18.5 t/s |
| Radeon RX 6900 XT AMD | 16 GB | Partial offload 36/36 | n/a | 59.4 GB | 18.4 t/s |
| Radeon RX 6800 AMD | 16 GB | Partial offload 36/36 | n/a | 59.4 GB | 18.4 t/s |
| Arc A770 16GB Intel | 16 GB | Partial offload 36/36 | n/a | 59.4 GB | 18.4 t/s |
| GeForce RTX 5070 NVIDIA | 12 GB | Partial offload 36/36 | n/a | 59.4 GB | 18.1 t/s |
| GeForce RTX 3080 10GB NVIDIA | 10 GB | Partial offload 36/36 | n/a | 59.4 GB | 17.7 t/s |
| GeForce RTX 2080 Ti NVIDIA | 11 GB | Partial offload 36/36 | n/a | 59.4 GB | 17.7 t/s |
| GeForce RTX 4070 Ti NVIDIA | 12 GB | Partial offload 36/36 | n/a | 59.4 GB | 17.6 t/s |
| GeForce RTX 4070 SUPER NVIDIA | 12 GB | Partial offload 36/36 | n/a | 59.4 GB | 17.6 t/s |
| GeForce RTX 4070 NVIDIA | 12 GB | Partial offload 36/36 | n/a | 59.4 GB | 17.6 t/s |
| GeForce GTX 1080 Ti NVIDIA | 11 GB | Partial offload 36/36 | n/a | 59.4 GB | 17.3 t/s |
| GeForce RTX 4080 Laptop NVIDIA | 12 GB | Partial offload 36/36 | n/a | 59.4 GB | 17.3 t/s |
| GeForce RTX 4060 Ti 16GB NVIDIA | 16 GB | Partial offload 36/36 | n/a | 59.4 GB | 17.1 t/s |
| GeForce RTX 3070 Ti NVIDIA | 8 GB | Partial offload 36/36 | n/a | 59.4 GB | 16.9 t/s |
| GeForce RTX 3060 12GB NVIDIA | 12 GB | Partial offload 36/36 | n/a | 59.4 GB | 16.9 t/s |
| Radeon RX 7700 XT AMD | 12 GB | Partial offload 36/36 | n/a | 59.4 GB | 16.9 t/s |
| Arc B580 Intel | 12 GB | Partial offload 36/36 | n/a | 59.4 GB | 16.8 t/s |
| GeForce RTX 2060 12GB NVIDIA | 12 GB | Partial offload 36/36 | n/a | 59.4 GB | 16.7 t/s |
| Radeon RX 6700 XT AMD | 12 GB | Partial offload 36/36 | n/a | 59.4 GB | 16.6 t/s |
| Radeon RX 7600 XT AMD | 16 GB | Partial offload 36/36 | n/a | 59.4 GB | 16.5 t/s |
| GeForce RTX 5060 Ti 8GB NVIDIA | 8 GB | Partial offload 36/36 | n/a | 59.4 GB | 16.4 t/s |
Can my GPU run GPT OSS 120B?
A direct answer for popular devices, each with the settings we would use.
How it compares
Other models in the same family or of a similar size, on a 12 GB card and a 24 GB card, each at the best quantization that fits.
| Model | Weights at 4 bits | GeForce RTX 3060 12GB | GeForce RTX 4090 |
|---|---|---|---|
| GPT OSS 120B | 59.0 GB | 16.9 t/s at MXFP4 | 23.9 t/s at MXFP4 |
| GPT OSS 20B | 11.3 GB | 65.7 t/s at MXFP4 | 180.1 t/s at MXFP4 |
| Qwen3.5 122B A10B | 70.3 GB | 12.6 t/s at Q4_K_M | 17.7 t/s at Q4_K_M |
| Qwen3 Coder Next | 44.8 GB | 33.2 t/s at Q4_K_M | 56.4 t/s at Q4_K_M |
| Qwen3 Next 80B A3B Instruct | 45.7 GB | 32.7 t/s at Q4_K_M | 54.6 t/s at Q4_K_M |
| GLM 4.5 Air | 62.1 GB | does not fit well | 10.1 t/s at Q4_K_M |
What to buy to run GPT OSS 120B
The cheapest hardware that runs it at a quantization worth using and a speed you would not resent, at an 8k context window.
Cheapest that works
GeForce GTX 1660 SUPER
$110 used
MXFP4 · uses 59.4 GB of its 6 GB · about 15.4 tokens/s
Best value
Ryzen AI Max+ 395 96GB
$1,999 whole machine
MXFP4 · about 26.8 tokens/s · 256 GB/s
If it just has to run
GeForce GTX 1650
$100 used
MXFP4 · about 12.6 tokens/s. Cheaper, and slower to answer.
| Device | Price | Memory | Runs it at | Speed | |
|---|---|---|---|---|---|
| GeForce GTX 1660 SUPER NVIDIA | $110 used | 6 GB | MXFP4 | 15.4 t/s | Check price |
| GeForce GTX 1080 Ti NVIDIA | $150 used | 11 GB | MXFP4 | 17.3 t/s | Check price |
| GeForce RTX 2060 12GB NVIDIA | $160 used | 12 GB | MXFP4 | 16.7 t/s | Check price |
| GeForce RTX 2080 Ti NVIDIA | $250 used | 11 GB | MXFP4 | 17.7 t/s | Check price |
| Arc A770 16GB Intel | $280 used | 16 GB | MXFP4 | 18.4 t/s | Check price |
| GeForce RTX 3080 12GB NVIDIA | $400 used | 12 GB | MXFP4 | 18.6 t/s | Check price |
| Radeon RX 7800 XT AMD | $480 | 16 GB | MXFP4 | 18.9 t/s | Check price |
| Radeon RX 9070 AMD | $560 | 16 GB | MXFP4 | 19 t/s | Check price |
Indicative prices reviewed 2026-09-01; used prices are marketplace typical. Options that cost more than a cheaper one with no more memory and no more speed are hidden. Price links are Amazon affiliate links. Change the standard, the context window or the budget in the buying tool.
Where it sits in the catalog
Source: openai/gpt-oss-120b on Hugging Face. Downloaded 5.4M times in the last month. Published 2025-08-04. Architecture figures are read from the repository's own configuration file, so they move when the model does.
Questions people ask about GPT OSS 120B
How much VRAM does GPT OSS 120B need?
At MXFP4 with an 8k context window, GPT OSS 120B needs about 59.4 GB: 58.74 GB of weights, 0.29 GB of KV cache and 0.37 GB of overhead. At the highest-quality file it needs about 59.4 GB.
What is the cheapest GPU that runs GPT OSS 120B?
The smallest consumer device in our catalog that fits it is the Apple M3 Ultra 96GB with 96 GB, running it at MXFP4 at roughly 102.5 tokens per second.
How much does a longer context window add for GPT OSS 120B?
On the memory side, a 32k window adds 1.1 GB of KV cache on top of the weights, and the cache grows more slowly than a conventional model because of its attention design. The longest window the model supports natively is 128k tokens.