Let’s talk quants of Gemma and Qwen - 16 vs Q8 vs Q4 - any experiences?

<!-- 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>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>so as it says, I am not a FE eng, but want to do some light FE work<br /> I don't need the smartest model but need to get some work done.<br /> I

<!-- SC_OFF --><div class="md"><p>Have Qwen 3.6 27B and Qwen 3.6 35B basically made most of the older ~30B models irrelevant?</p> <p>They seem to beat stuff like Qwen coder 30B,

<!-- SC_OFF --><div class="md"><p>I seen this one mentioned but it was a source from about 14 months ago.</p> <p>In the age of the Qwen 3.6 and Gemma 4- is there still a use for

<!-- SC_OFF --><div class="md"><p><a href="https://huggingface.co/unsloth/gemma-4-E2B-it-GGUF">https://huggingface.co/unsloth/gemma-4-E2B-it-GGUF</a></p> <p><a

<!-- SC_OFF --><div class="md"><p>What are your experiences with Gemma 4's QAT versions compared to their regular ones? So far I have mostly heard about regressions, but if you

<!-- SC_OFF --><div class="md"><p>Looking for your experiences, your speeds, and your configs. </p> <p>I myself am getting an abysmal 22t/s with the q4, max context, llama.cpp,