Qwen 3.6 27B on Strix Halo 128GB: any experiences?

<!-- SC_OFF --><div class="md"><p>I dont have good experience running q4_k_m, the difference to q6 is &quot;a few errors an hour&quot; to &quot; a few errors every couple of

<!-- SC_OFF --><div class="md"><p>With practically all of LocalLlama glazing Qwen 3.5/3.6 for it's coding skills. Along with the fact that Alibaba themselves are focusing on

<!-- SC_OFF --><div class="md"><p>Really loving Qwen 27b more than any other llm from when I can remember. It works so well. Having 48gb vram can anyone recommend any other

<!-- SC_OFF --><div class="md"><p>I'm genuinely unsure if I'm doing something wrong, or if the model is just worse, despite the community enthusiasm. I tend to use it with pi

<!-- SC_OFF --><div class="md"><p>I have a 5080 + 64gb of ram. What model would be as intelligent as possible while still running decent enough on my specs?</p> </div><!-- SC_ON

<!-- SC_OFF --><div class="md"><p>Has anyone gotten KV cache offloading working with Qwen on vLLM? No matter what configuration I try, I get errors and it crashes. I saw an old

<!-- SC_OFF --><div class="md"><p>Do we.know whether the 27B model will ship with a DFlash or MTP head? It's super exciting, but since 35B-A3B is my daily driver, 27B will crawl —

<!-- SC_OFF --><div class="md"><p>So in response to the Great Token Reconning of 2026, I decided to try out Qwen 3.6 as a daily driver, and although it's only been about a day, I