LocalClaw/LLM/GPT-OSS (20B)
Open-weight local LLM

GPT-OSS (20B)

OpenAI open-weight reasoning model. Strong general, coding and tool-use capabilities under Apache 2.0. Practical on 16GB+ machines.

16 GB sweet spot 16 GB RAM MXFP4 General AI assistant
Run with LocalClaw Compare all models
Parameters
20B
Minimum RAM
16 GB
Model size
12 GB
Quantization
MXFP4

Can GPT-OSS (20B) run locally?

GPT-OSS (20B) is a practical pick for 16 GB machines, especially with MXFP4 quantization.

Search for gpt-oss-20b in LM Studio or another GGUF-compatible runtime.

openai/gpt-oss-20b
chatcodereasoningpowergeneral

Install path

01
Check RAM fitMinimum 16 GB RAM. Start with the MXFP4 quant.
02
Load the modelSearch gpt-oss-20b in LM Studio.
03
Control locallyUse LocalClaw to manage models, agents, chat, channels and scheduled OpenClaw work.

Strengths

  • OpenAI open-weight reasoning model
  • Apache 2.0 license
  • Runs on 16GB-class local machines
  • 128K context
  • Strong tool use and coding

Limitations

  • Smaller than the 120B variant
  • MXFP4 runtime support is newer than standard GGUF quant families
  • Requires harmony-compatible prompting for best results

Best use cases

  • General AI assistant
  • Coding
  • Reasoning
  • Tool-use agents
  • Research comparison

Capability profile

speed
5
quality
8
coding
8
reasoning
9

Technical notes

Developer
OpenAI
License
Apache 2.0
Context window
131,072 tokens
Architecture
Sparse MoE Transformer with 21B total parameters, 3.6B active parameters and 128K context

This model fits these next steps

Hardware fit is based on LocalClaw's RAM tier, model size and quantization metadata. Always leave memory headroom for your OS and runtime.

Starter desktopMac mini M4 16GBPortable fitMacBook Air M4 16GBBest 16GB models16GB RAM guide

Similar models to compare

GPT-OSS (120B) 117B (5.1B active)Qwen 3 (14B) 14BPhi-4 (14B) 14BMistral Small (24B) 24B

Where to go next