I don't have a super high memory gpu, 16gb. But I got it to run `huihui-ai/Huihui-Qwen3.6-27B-abliterated-MTP-GGUF:Q3_K ` in llama.cpp and connected OpenCode to it. It is slow, but it is getting some work done.
Yah, in many ways it under performs. Like it's performance is similar to the previous generation card 7900 XTX,. And the vram is slower. But it does have 32GB of vram. And it's performance is similar to a 5070.
Modified by: Tylor on 2026-06-29 08:53:43 PM
Last edited by ViperDiscord on Tue Jun 30, 2026 1:53 am, edited 1 time in total.
I have the components for another machine, but with half the power of my main system. Hard to match when you run your system until they start to cause glitches from the age.