Can I run gpt-oss-120b locally?
Of the setups on this page, only the 2 two-card PCs run it, and only slowly: gpt-oss-120b needs about 64 GB of graphics memory, so part of it has to run from normal memory. A single card does not have enough graphics memory on its own. In a PC with enough normal memory it also runs slowly on one card.
The cheapest machine on our list that runs it: Mac Studio M4 Max 128GB, $3,683€3,290. price unverified
Our price is for a used Mac. These shops mostly sell new ones, so their price can be higher.
EU shops · In the US? Switch the currency to USD.
US shops · In Europe? Switch the currency to EUR.
We don't give a speed for this model yet. This model only uses part of itself for each word (a mixture of experts), and our speed estimate does not handle that yet.
Which graphics cards run it
Seven popular graphics cards and two PCs with two cards each, at the standard quality (Q4). Click one for the full answer: memory, speed and price.
Runs slowly2
Part of the model spills into normal memory, which is much slower.
Doesn't fit7
Not enough graphics memory: it needs about 64.5 GB. In a PC with enough normal memory it runs slowly: Budget PC with RTX 3060 12GB, $1,305€1,357. price unverified
- RTX 3060 12GBhas 12 GB
- RTX 4060 Ti 16GBhas 16 GB
- RTX 3090has 24 GB
- RTX 4090has 24 GB
- RTX 5080has 16 GB
- RTX 5090has 32 GB
- RX 7900 XTXhas 24 GB
Cheapest hardware for each size
From all 36 graphics cards, PCs and Macs on our list. "Runs slowly" means part of the model runs from normal memory: cheaper, but much slower. A smaller download needs less memory and answers a little less well.
- Standard quality (MXFP4)Mac Studio M4 Max 128GB, $3,683€3,290 price unverified
- Standard quality (MXFP4), runs slowlyBudget PC with RTX 3060 12GB, $1,305€1,357 price unverified
Prices excl. shipping. Euro prices incl. VAT. Dollar prices are converted from our euro prices without EU VAT, before US sales tax.
Prices moved: see Building a local LLM rig in the RAMageddon.
About gpt-oss-120b
117B total, 5.1B active. Parameter count is an estimate. More parameters usually means a smarter model that needs more memory.
- Licence:
- Apache-2.0 source
- Longest conversation:
- about 197 pages of text (131,072 tokens)
- Input:
- text
Technical notes
OpenAI's larger open model for reasoning, code and tool use. It ships in 4-bit, so the official file is about 63 GB: no single graphics card on our list holds it, but a 96 GB or 128 GB Mac or a 128 GB mini PC does. Even the full-quality file is only 65.4 GB, so our higher quality levels use that file's size. Lower-quality versions we checked keep that 4-bit core and save under 1 GB.
Memory at each quality level
Each download of gpt-oss-120b is a different size. The total adds the file, the memory a conversation of about 12 pages of text (8,192 tokens) takes, and 0.8 GB of working space. A measured file size is labelled; the rest is our estimate. The last column is the cheapest machine that holds the total.
Swipe the table sideways to see every column, up to the cheapest machine.
| Quality (file) | File size | Conversation | Working space | Total | Cheapest machine that holds it |
|---|---|---|---|---|---|
| F16 | 65.37 GBmeasured GGUF | 0.3 GB | 0.8 GB | 66.5 GB | Mac Studio M4 Max 128GB, $3,683€3,290 price unverified |
| MXFP4 | 63.39 GBmeasured GGUF | 0.3 GB | 0.8 GB | 64.5 GB | Mac Studio M4 Max 128GB, $3,683€3,290 price unverified |
speed: not estimated for MoE yet.
Waiting for prices to drop?
Speeds and memory are our estimate from the hardware specs, not a measured test. Prices are dated and may change. How we calculate