GPT-OSS Models

Browse and run open-source AI models locally on your own hardware — no API keys, no cloud, no usage limits. Fully private, works offline, and free forever.

GPT-OSS running in Atomic Chat
Qwen3.6 27B MTP
27B
No models found
Reasoning
Code
Multilingual
Web
Vision
Thinking
Tools
Audio
Visuals
Embedding
gpt-oss-120b

gpt-oss-120b is OpenAI's open-weight MoE: 117B parameters, 5.1B active, sized for a single 80 GB GPU. Run it locally, free, with Atomic Chat.

Tools
Thinking
Reasoning
Code
gpt-oss-20b

gpt-oss-20b is OpenAI's open-weight 21B MoE model with 3.6B active parameters, with its MoE weights post-trained in MXFP4 to run within 16 GB of memory.

Tools
Thinking
Reasoning
Code