GPT-OSS 120B
MoEToolsPopularOpenAI · 117B (5.1B active) · Mixture of Experts
OpenAI's flagship open-weight MoE — 52.6% SWE-bench
4.4M
4.5K
2025-08
128K tokens
Use cases
Your hardware
16 GB Reference Desktop
Quantization options
What is quantization?| Quant | Bits | VRAM | Disk | Quality | Status |
|---|---|---|---|---|---|
| Q2_K | 2 | 38 GB | 35.8 GB | low | Barely runs · ~1 t/s |
| Q3_K_M | 3 | 52.9 GB | 50.1 GB | moderate | Too heavy · ~1 t/s |
| Q4_K_M | 4 | 60.4 GB | 57.2 GB | good | Too heavy · ~1 t/s |
| Q5_K_M | 5 | 75.4 GB | 71.5 GB | good | Too heavy · ~1 t/s |
| Q6_K | 6 | 90.4 GB | 85.8 GB | excellent | Too heavy · ~1 t/s |
| Q8_0 | 8 | 120.4 GB | 114.4 GB | excellent | Too heavy · ~1 t/s |
| F16 | 16 | 240.2 GB | 228.8 GB | lossless | Too heavy · ~1 t/s |
About this model
GPT-OSS 120B (GPT-oss-120b) is a 117B sparse mixture-of-experts (MoE) model activating 5.1B active per token from OpenAI in the GPT-OSS family, designed primarily for chat, reasoning, code workflows. Supports structured tool calling; use Q4_K_M or higher when strict JSON schema adherence matters. Supports up to 128K context tokens; start at 4K–8K with Q8_0 KV cache on tight GPUs.
Hardware planning baselines start at 65.4 GB minimum memory and 109 GB recommended memory. At Q4_K_M, weights plus baseline runtime overhead use roughly 60.4 GB of RAM/VRAM and 57.2 GB of disk space. At 117B, GPT-OSS 120B belongs firmly in the workstation and multi-GPU tier. Running it comfortably requires 109 GB of total memory across dual RTX GPUs, high-memory Apple Silicon (Mac Studio Ultra/Max), or MoE hybrid CPU-RAM offloading—mobile devices and standard 16 GB laptops are too constrained for practical speeds.
Deploy GPT-OSS 120B locally via Ollama (`gpt-oss:120b`), LM Studio, or Ollama with Flash Attention enabled. Under Apache License 2.0, review official weight cards before commercial redistribution.
Highlights
- Local Fit Baseline65.4 GB minimum and 109 GB recommended RAM/VRAM.
- Primary QuantQ4_K_M targets 60.4 GB memory and 57.2 GB disk.
- Context & Capabilities128K max context · Sparse MoE · Tool calling.
System requirements by device
Memory planning targets 65.4 GB minimum RAM/VRAM and 109 GB recommended memory. The baseline quant is Q4_K_M, estimated at 60.4 GB memory and 57.2 GB disk. Actual runtime usage shifts with context length, KV cache precision, and OS background overhead.
| Device | Notes |
|---|---|
| Mac (Apple silicon) | Requires a 48 GB–192 GB Mac Studio or MacBook Pro Max/Ultra with enough unified wired budget for 60.4 GB. |
| Windows / Linux GPU | Use dual/multi-GPU CUDA splitting (--tensor-split) or MoE CPU expert offloading (-ot) paired with 32–128 GB DDR5 system RAM. |
| Phones & Compact Edge Boards | Not practical on phones (65.4 GB min RAM exceeds Android LMK and iOS Jetsam limits). Pick a 2B–4B model on mobile. |
Start with a short session and leave memory for the operating system. A file fitting on disk does not guarantee that it will load in memory.
Estimated memory by quantization
These are rough estimates, not tested minimum requirements. Check the actual file and runtime documentation; context, offloading and operating-system overhead can change the memory needed.
| Quant Tier | Estimated memory | Disk Size | Quality |
|---|---|---|---|
| Q2_K | 38.0 GB | 35.8 GB | low |
| Q3_K_M | 52.9 GB | 50.1 GB | moderate |
| Q4_K_M | 60.4 GB | 57.2 GB | good |
| Q5_K_M | 75.4 GB | 71.5 GB | good |
| Q6_K | 90.4 GB | 85.8 GB | excellent |
| Q8_0 | 120.4 GB | 114.4 GB | excellent |
| F16 | 240.2 GB | 228.8 GB | lossless |
How to run GPT-OSS 120B on your device
Start with a short prompt and the smallest quant that still gives the quality you need. A clean first run is more useful than loading a larger file with no memory headroom.
Run GPT-OSS 120B on Multi-GPU or High-RAM Desktop
- Plan for 65.4 GB minimum combined memory and 109 GB recommended memory for Q4_K_M.
- Use llama.cpp `-ot ".ffn_.*_exps.=CPU"` or `--tensor-split` across multiple GPUs so attention and KV cache remain in fast VRAM.
- Enable Flash Attention (`-fa on`) and Q8_0 KV cache (`-ctk q8_0 -ctv q8_0`) to prevent context buffers from exceeding GPU headroom.
- Avoid attempting to run workstation-scale weights on smartphones or 8 GB laptops; use a 2B–8B checkpoint on portable devices.
Run GPT-OSS 120B on High-Memory Apple Silicon Mac
- Verify you are on a 48 GB–192 GB Mac Studio or MacBook Pro Max/Ultra capable of holding 60.4 GB in wired unified memory.
- Launch via `llama-server` or LM Studio with MLX/Metal enabled and set initial context conservatively to 4K–8K.
- Monitor Activity Monitor Memory Pressure to confirm zero SSD swap thrashing during token generation.
- Step down to a lighter quantization (e.g., Q3_K_M or IQ4_XS) if total memory utilization exceeds 75%.
GPT-OSS 120B vs alternatives
| Model | Params | Context | Min RAM | License |
|---|---|---|---|---|
| GPT-OSS 120B | 117B (5.1B active) | 128K | 65.4 GB | Apache 2.0 |
| GPT-OSS 20B | 21B (3.6B active) | 128K | 11.7 GB | Apache 2.0 |
| Mistral Small 4 119B | 119B (6.5B active) | 256K | 66.5 GB | Apache 2.0 |
| Qwen 3.5 122B-A10B | 122B (10B active) | 256K | 68.2 GB | Apache 2.0 |
| Devstral 2 123B | 123B | 256K | 68.7 GB | MRL |
| Llama 4 Scout 17B | 109B (17B active) | 128K | 60.9 GB | Llama 4 Community |
Use this table to choose the smallest nearby model that still matches the job. A related model with lower RAM can be the better local choice when the current page fits only at the edge of available memory.
Frequently asked questions
- What is GPT-OSS 120B?
- GPT-OSS 120B is a 117B sparse mixture-of-experts model (5.1B active active per token) from OpenAI, supporting up to 128K tokens with 65.4 GB minimum RAM and 109 GB recommended RAM.
- What is the best way to run GPT-oss-120b locally?
- Choose the smallest quant that matches the task, test one short prompt, then compare RAM, context, and nearby models before making GPT-OSS 120B your default local model.
- How much VRAM or RAM does GPT-OSS 120B require?
- At Q4_K_M, GPT-OSS 120B occupies roughly 60.4 GB of active memory and 57.2 GB on disk. Add 2–6 GB of spare headroom for KV cache buffers, runtime scratch space, and OS apps (109 GB recommended).
- What is the best way to run GPT-OSS 120B locally?
- Use `ollama run gpt-oss:120b` for a quick CLI/API setup, or load Q4_K_M in LM Studio or llama.cpp with Flash Attention enabled when you want explicit KV cache and layer control.
- Can GPT-OSS 120B run on a consumer laptop or smartphone?
- No smartphone or 8–16 GB laptop can hold GPT-OSS 120B cleanly in local memory. It targets multi-GPU NVIDIA workstations, high-RAM Mac Studios, or MoE CPU-RAM hybrid rigs with at least 65.4 GB total memory.
- How does GPT-OSS 120B compare to GPT-OSS 20B?
- GPT-OSS 20B (21B, 11.7 GB min RAM) sits closest in size. Compare active parameters, context ceiling (128K vs 128K), and license terms (Apache 2.0) before switching.