OpenAI - Mixture of experts - verified
Run gpt-oss-120b locally
gpt-oss-120b is a 116.8B mixture-of-experts (5.1B active) model from OpenAI with a 131K-token context window, released under Apache 2.0 (commercial use is allowed); the smallest listed download is 62.3 GB, the MXFP4-Q4 MLX build from mlx-community/gpt-oss-120b-MXFP4-Q4.
Official weights: openai/gpt-oss-120b
- Total params
- 116.8B
- Active params
- 5.1B
- Context
- 131K
- Smallest download
- 62.3 GB
Fits in
Smallest download at most 85% of memoryDownloads
The official weights and every quantization listed for this model, with file sizes from Hugging Face.
| Download | Format | Size | Repo |
|---|---|---|---|
Official weights | MXFP4 | 65.2 GB | openai/gpt-oss-120b |
MXFP4 | GGUF | 63.4 GB | ggml-org/gpt-oss-120b-GGUF |
Q4_K_M | GGUF | 62.8 GB | unsloth/gpt-oss-120b-GGUF |
Q8_0 | GGUF | 63.4 GB | unsloth/gpt-oss-120b-GGUF |
MXFP4-Q4Smallest listed | MLX | 62.3 GB | mlx-community/gpt-oss-120b-MXFP4-Q4 |
GB is decimal: 1 GB = 10^9 bytes. Add memory for the context and the operating system on top of the file size.
Run it
Copied verbatim from the page linked under each command. Server commands from model cards often assume a multi-GPU node; check the flags before you run them.
Transformers
pip install -U transformers kernels torchNot sure which runtime to use? Compare local LLM runtimes by platform, API and features.
License
- License
- Apache 2.0
apache-2.0 - Commercial use
- Commercial use: YesCommercial use is allowed.
- Gated download
- No. The files download directly.
Worth knowing
Must be prompted in OpenAI's harmony response format, and its MXFP4 MoE weights let it fit on a single 80GB GPU.
Released (per the announcement).
FAQ
How big is the gpt-oss-120b download?
The official MXFP4 weights in openai/gpt-oss-120b are 65.2 GB. The smallest listed download is 62.3 GB, the MXFP4-Q4 MLX build from mlx-community/gpt-oss-120b-MXFP4-Q4. Plan for more memory than the file size: the context (KV cache) and the operating system need room too.
Can I use gpt-oss-120b commercially?
gpt-oss-120b is released under Apache 2.0, and commercial use is allowed. The license text is at https://huggingface.co/openai/gpt-oss-120b/blob/main/LICENSE.
How much memory do I need for gpt-oss-120b?
By this directory's rule of thumb it fits in 128 GB and up: the smallest listed download, 62.3 GB, is at most 85% of 128 GB, which leaves room for the context and the operating system.
How do I run gpt-oss-120b with Ollama?
Run "ollama run gpt-oss:120b", as shown at https://huggingface.co/openai/gpt-oss-120b. The Ollama library lists the gpt-oss:120b tag at 65GB.
What is the context length of gpt-oss-120b?
131,072 tokens. A longer context needs more memory for the KV cache, so you can run with a shorter one when memory is tight.
Sources
Last verified from the pages below.
- HF API - huggingface.co
- HF file tree - huggingface.co
- Model card - huggingface.co
- config.json - huggingface.co
- Ollama gpt-oss - ollama.com
- OpenAI: Introducing gpt-oss - openai.com
Newsletter
New open weights, every week
One email per week. New model releases, open-weight drops and what they mean for what you build.
Keep exploring
Best local AI models
Our curated picks by hardware tier, with benchmarks and the reasoning.
Model API pricing and specs
First-party prices per 1M tokens when you would rather call an API.
AI benchmark leaderboards
Sourced scores, for when you need to compare quality.
Setup guides
Point your coding tools at the model you picked.