OpenAI GPT OSS 20B on a 16GB Mac — Ollama Setup, Demo & Performance Tips
#gptoss #ollama #localllm
#openai #openweight #lmstudio #huggingface #ai #agents #reasoning
Welcome back to the channel! This week was a wild week for AI. OpenAI dropped 2 open weights model, and one of them - they said, can run on your local 16GB RAM, so here we go.
In this tutorial, we’ll spin up GPT-OSS 20B with Ollama on a 16GB Mac, run a quick QA demo, share performance tips, and an honest review. Does it really can perform complex task on 16GB machine? If you’re into local AI, agentic workflows, or just curious how far a 16GB machine can go—this is for you.
🎥 In this video, we cover…
Watch on YouTube ↗
(saves to browser)
Chapters (10)
Intro — What is GPT-OSS & why local?
0:29
Multiple ways to run GPT-OSS
2:04
Ollama requirements (memory, disk, expectations)
3:04
Pull & run gpt-oss:20b (terminal demo)
3:19
Ollama GUI
4:03
Crashing my MAC
5:09
Changing config to make GPT-OSS work!
7:34
Running Ollama again with lighter GPT-OSS after config
7:40
Result
8:40
Honest Thought - Can it run? Yes! Do I have the patience? No
DeepCamp AI