Re:Ask HN: Who uses open LLMs and coding assistants locally?...
by mapontosevenths on Friday October 31, 2025 @03:08PM
(#45775557)
I run GPT OSS 120b locally on my Asus GX10 (their version of the DGX Spark), and it's great. It usually takes a little less than 60 seconds before it starts answering, but once it's started it generates the tokens fast enough for my needs.
I use the 20b version when I'm in a hurry or the task is trivial to speed things up a bit.
I use the 20b version when I'm in a hurry or the task is trivial to speed things up a bit.