OpenAI Just Went Open-Source — FULL gpt-oss 20B & 120B Testing!

Timestamps:

00:00 - Intro
01:33 - Technical Look
05:10 - Local Test Info
05:45 - Browser OS Test (120B)
06:15 - System Specs
10:27 - Code Fix Test
12:50 - Browser OS Test (20B)
17:50 - Python Game Test (20B)
21:08 - Python Game Test (120B)
23:45 - Website Generation Test (20B)
25:50 - Website Generation Test (120B)
28:51 - COT & Roleplay Testing (20B)
31:31 - COT & Roleplay Testing (120B)
32:28 - Closing Thoughts

AI Consulting/Integration: https://bijanbowen.com
Discord:   / discord  

In this video, we take a look at the newly released gpt-oss family from OpenAI — their first open-source LLM release since GPT-2. With two models available (20B and 120B), this release marks a major shift in OpenAI’s approach to community-accessible models.

We begin with a technical overview of both models, then move into full local testing using Ollama. Throughout the video, we compare the performance of the 20B and 120B versions across a variety of tasks — including browser-based OS generation, Python game creation, bug fixing, website builds, and roleplay + chain-of-thought reasoning.

Hugging Face: https://huggingface.co/openai