EDGE OSOPENAI · 117B PARAMETERS
gpt-oss-120b hardware and VRAM requirements
OpenAI’s larger open-weight reasoning model. OpenAI says it fits on a single 80 GB GPU such as an NVIDIA H100 or AMD MI300X. gpt-oss-120b is a mixture-of-experts model with 5.1B of its 117B parameters work on each token. At MXFP4 it needs about 70 GB of GPU memory for one user, and about 75 GB to serve a team of 16 with 8K context each.
Total parameters
117B
Active per token
5.1B
Context length
128K
Released as
MXFP4 (native)
Minimum VRAM
70 GB at MXFP4
For a team of 16
75 GB at MXFP4
VRAM needed to run gpt-oss-120b
GPU memory by precision, number of concurrent users and context length. Each user needs their own KV cache, so memory grows with users and context.
| Precision | Weights only | 1 user, 8K | 16 users, 8K | 16 users, 32K | LoRA fine-tune |
|---|---|---|---|---|---|
| MXFP4 (native) Released | 66 GB | 70 GB | 75 GB | 89 GB | 89 GB |
Full fine-tuning in mixed precision needs about 1.89 TB: weights, gradients and optimizer states at roughly 16 bytes per parameter. LoRA trains small adapter layers on a frozen copy of the model instead, so it needs far less.
GPUs that can run gpt-oss-120b on-premise
For a team of 16 using gpt-oss-120b at MXFP4 with 8K context each (about 75 GB), smallest fit first.
NVIDIA H100 NVL
94 GB per GPU · 9.8 GB to spare.
NVIDIA RTX PRO 6000 Blackwell
96 GB per GPU · 12 GB to spare.
NVIDIA DGX Spark
128 GB per system · 34 GB to spare.
NVIDIA H200 NVL
141 GB per GPU · 52 GB to spare.
AMD Instinct MI300X
192 GB per GPU · 98 GB to spare.
AMD Instinct MI325X
256 GB per GPU · 156 GB to spare.
NVIDIA DGX B200
1.44 TB per system · 1.22 TB to spare.
NVIDIA L40S
48 GB per GPU · Split across 2 GPUs in one server, 12 GB to spare.
Some of these setups need several GPUs working as one. Buying the GPUs is the easy part. Vanik Edge OS splits the model across them, keeps it running if a server fails and manages it all from one console. See how Edge OS clusters GPU servers
Weighing two of these? Compare AI hardware side by side, for example H100 vs H200, RTX PRO 6000 vs RTX 5090 or DGX Spark vs Mac Studio.
Running gpt-oss-120b for your team?
We will size the GPU server for your users and data, and set up gpt-oss-120b on it inside your own building. Free, no obligation.
gpt-oss-120b at a glance
Compare gpt-oss-120b with other models
FROM SIZING TO RUNNING
Sizing the hardware is the easy part.
Getting several GPUs to work as one system for a whole company is where most projects slow down. That is the part Vanik Edge OS does for you.
STEP 01 · THIS TOOL
Size it
Work out the GPU memory your models need and which GPUs or servers can hold them.
STEP 02 · VANIK EDGE OS
Cluster it
Edge OS joins your GPU servers into one cluster, splits a large model across them and keeps it running if a server fails. NVIDIA and AMD servers work side by side.
STEP 03 · VANIK EDGE OS
Run it
Your team gets AI apps with single sign-on and an audit trail, managed from one console, with your data staying inside your building.
gpt-oss-120b GPU questions
Can I run gpt-oss-120b locally?
Yes, on data-center GPUs. At MXFP4 it needs about 70 GB of GPU memory for one user, more than consumer graphics cards or laptops hold. A team of 16 fits on NVIDIA H100 NVL.
What are the gpt-oss-120b hardware requirements?
To run gpt-oss-120b locally you need about 70 GB of GPU memory at MXFP4 for one user (75 GB for a team of 16), system RAM of at least 66 GB so the weights can be loaded, around 66 GB of fast NVMe storage for the model files, and an inference engine such as vLLM or SGLang. A team of 16 fits on NVIDIA H100 NVL.
How much VRAM to run gpt-oss-120b?
gpt-oss-120b needs about 70 GB of GPU memory at MXFP4 for one user with 8K context: 66 GB for the weights plus the KV cache and runtime overhead. Serving 16 people at once with 8K context each needs about 75 GB.
Can gpt-oss-120b run on a single H100 80GB?
Yes, for light use: at MXFP4 with one user it needs about 70 GB, within the roughly 72 GB usable on an 80 GB H100.
How big is gpt-oss-120b?
gpt-oss-120b has 117B parameters in total, of which 5.1B work on each token. The weights take about 66 GB at MXFP4.
Is gpt-oss-120b multimodal?
No. gpt-oss-120b is a text-only model: it takes text in and produces text out.
Is gpt-oss-120b free?
Yes. OpenAI released the weights under the Apache 2.0 license, so you can download and run it at no cost and use it commercially. You only pay for the hardware it runs on.
Why does gpt-oss-120b need so much memory if only 5.1B parameters are active?
It is a mixture-of-experts model. Only 5.1B parameters work on each token, which keeps it fast, but a different set of experts is picked for every token, so all 117B parameters must stay loaded in GPU memory.
Can I fine-tune gpt-oss-120b?
LoRA fine-tuning at MXFP4 needs roughly 89 GB with 4K sequences; it trains small adapter layers on a frozen copy of the model. Full fine-tuning needs about 1.89 TB, which calls for a multi-GPU training cluster.
How are these numbers calculated?
Weights are parameters × bytes per parameter. The KV cache uses 36.0 KB per token per user. We add about 5% of the weights plus 1 GB for runtime overhead and keep 10% of each GPU free. Figures are planning estimates; real usage depends on the inference engine and settings.
VRAM for similar models
Specifications from Hugging Face: openai/gpt-oss-120b. Memory and bandwidth are manufacturer figures; GPU counts are rounded to 1, 2, 4 or 8 per server. All figures are planning estimates. Last checked November 10, 2026.
Run gpt-oss-120b inside your own building
Vanik Edge OS runs open-weight models on NVIDIA DGX, AMD Instinct and other GPU servers on your premises, so your data never leaves.