Hey fellow Llamas, your time is precious, so I'll keep it short (while trying to explain everything lol). TL;DR: 33-35B MoE on a 16 GB GPU. Qwen3.6 35B-A3B: 13.3 GiB (was ~20.5). Laguna XS.2 33B-A3B: 14.6 GiB (was 18.8). Both measured on an RTX 3090, both under 16 GiB. Only the active experts stay on the GPU. An A3B model routes to ~8 of 256 experts per token. Spark calibrates which experts your traffic hits and keeps those hot; the long tail lives in system RAM and is swapped in o...