can it run
Comfortably. 128GB of memory against a 33GB working set leaves headroom for context, the OS, and anything else you're running.
weights 26GB + KV cache 0.5GB at 4k context vs ~126GB usable → full. 26.5 GB needed of 126 GB usable — headroom for context
computed roofline: 256 GB/s × 0.60–0.85 efficiency window / 3.9GB (Q4_K_M) — bands, never points.
MoE: file is 26GB (that decides fit) but only ~1.05B of experts activate per token (~3.9GB streamed) — band below is the active-expert estimate, optimistic; community MoE rows run several× the naive bandwidth math
| model file | 26GB (GGUF Q4_K_M) |
| minimum memory | 33GB working set |
| this machine | 128GB LPDDR5X-8533 (~276 GB/s) |
| params | 8x7B MoE |
| license | apache-2.0 |
| price | $4099 |
easy run: ollama pull mixtral:8x7b
runs with: ollama · llama.cpp · koboldcpp · lm studio
→ ASUS ExpertCenter Pro ET900N G3 (GB300) — this model on the biggest machine