can it run
23.0 of 24 GB usable — keep context modest
24GB · RDNA3 · model file 22GB · 32B
computed roofline: 960 GB/s × 0.50–0.75 efficiency window / 22.0GB (Q4_K_M) — bands, never points.
Community check: ~44 tok/s · 30-33B Q4 · 9 runs · just above our band (25% over the high edge) — localmaxxing.com
→ Mac mini M5 Pro (64GB) — cheapest catalog machine that runs qwen2.5vl 32B fully ($1669)
→ rtx-3090 — the next step up from AMD Radeon RX 7900 XTX
→ every model the AMD Radeon RX 7900 XTX can run — full list
→explore the full catalog84 machines indexed · live prices · what each one can run