Agents / OpenAI
gpt-oss · 20B
An open-weight reasoning and tool-use model with a publisher-documented Ollama path for consumer hardware.
Verified sourceSource checked 9/23/2026
LOCALRENTED GPUOPEN WEIGHTS
At a glance
- Parameters
- 21B total · 3.6B active
- Architecture
- Mixture of experts
- License
- Apache 2.0
- Disk space
- 14 GB
- Software
- Ollama, vLLM, Transformers
Sources and files
Publisher model card and Ollama commands Ollama package sizeWill it run on your machine?
Save your machine to see a personalized rating and its reasoning.
Add my hardwareBest for
- Run local reasoning with adjustable effort.
- Build a private structured-data extraction or tool-use process.
Tradeoffs
- Use a runtime that handles the model’s Harmony format; direct raw prompting can behave incorrectly.
- The 14 GB package size does not include all runtime and context memory.
Ways to run it
Ollama · Windows, macOS, Linux
The publisher documents this command; Ollama lists a 14 GB package. It uses the model’s Harmony format.
ollama run gpt-oss:20bExplore its uses
REASONINGAGENTSTOOLSSTRUCTURED OUTPUT
Get it running
Extract structured data with a local model
Use gpt-oss 20B through Ollama to turn a short text sample into validated JSON fields, then check the output against the source.
View workflowKeep exploring
Agents
Z.aiGLM-5.3-Flash
A multimodal model for coding and agent tasks, with 18B active parameters from a 320B total model.
MODEL SIZE320B total · 18B active
Agents
Moonshot AIKimi K3
An open-weight multimodal model for long-running coding, reasoning, and knowledge work.
MODEL SIZE2.8T total · 104B active
Agents
Google DeepMindGemma 4 · 12B
A compact multimodal model that understands text, images, audio, and video for local assistants.
MODEL SIZE12B