MTP with Dual 3090's on Qwen 27B

<!-- SC_OFF --><div class="md"><p>Hey, </p> <p>Tldr: use local models like qwen 3.5 quantized with proprietary models for fire and forget work. Local model doing the grunt work.

<!-- SC_OFF --><div class="md"><p>I'm hoping you guys could help me here. Looking at the price of things I can get two 5060 16gb cards for about $1100 new giving me 32gb of vram

<!-- SC_OFF --><div class="md"><p>Hardware-----------------<br /> 4070 12gb - gen 4.0 x16<br /> v100 16gb - gen 3.0 x4</p> <p>32gb 3200mhz ram</p> <p>it5-12400f<br /> ssd read -

<!-- SC_OFF --><div class="md"><p>I'm currently choosing between IQ3_XXS Qwen-3.8-Flash and Q8_0 27B.</p> <p>This month I don't have anything complex enough to justify either's

<!-- SC_OFF --><div class="md"><p>PR #22673 (commit 4f13cb7) landed MTP speculative decoding in mainline llama.cpp on May 16. I tested it on two separate rigs.</p> <p>Qwen3.6 27B,

<table> <tr><td> <a href="https://www.reddit.com/r/LocalLLaMA/comments/1s6g67w/x13_dual_xeon_silver_4415_1_tb_ram_4_x_nvidia/"> <img alt="X13 + Dual Xeon Silver 4415 + 1 TB RAM +

<!-- SC_OFF --><div class="md"><p>I could fit the Q3 model not sure if it's worth over 27B ? </p> </div><!-- SC_ON --> &#32; submitted by &#32; <a

<!-- 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 —