OpenAI just dropped their first open-weight models since GPT-2. GPT-OSS-120B and GPT-OSS-20B, both under the Apache 2 license. In this video, I run the 20B model locally, put both through real coding and reasoning tests, and compare their performance to Qwen, DeepSeek, and OpenAI’s own o3/o4 Mini.
Are these the new go-to open models for devs? Let’s find out.
🔗 Relevant Links
Intro Blog Post: https://openai.com/index/introducing-gpt-oss/
Ollama: https://ollama.com
❤️ More about us
Radically better observability stack: https://betterstack.com/
Written tutorials: https://betterstack.com/community/
Example projects: https://github.com/BetterStackHQ
📱 Socials
Twitter: https://twitter.com/betterstackhq
Instagram: https://www.instagram.com/betterstackhq/
TikTok: https://www.tiktok.com/@betterstack
LinkedIn: https://www.linkedin.com/company/betterstack
📌 Chapters:
0:00 “Open”AI
0:40 GPT-OSS-20b
1:04 Running Locally
2:16 Coding Challenge
3:14 GPT-OSS-120b
4:25 Coding Test
7:08 Benchmarks
source