local intelligence

can it run

Can the RTX 3090 (24GB, used) run qwen3.8 27B?

TIGHT

Yes, but close to the edge. 24GB against a 21GB working set: it runs, with little room for long context windows or other processes.

~26–43 tok/s

weights 16.5GB + KV cache 2GB at 4k context vs ~24GB usable → full. 18.5 GB needed of 24 GB usable — headroom for context

computed roofline: 936 GB/s × 0.46–0.76 efficiency window / 16.5GB (Q4_K_M) — bands, never points.

Community check: ~26 tok/s · 20-27B Q4 · 20 runs · inside our computed band — localmaxxing.com

explore the full catalog84 machines indexed · live prices · what each one can run

Specs

model file16.5GB (GGUF Q4_K_M)
minimum memory21GB working set
this machine24GB GDDR6X, 936 GB/s
params27B
licenseapache-2.0
price$1100

Run it

model files: hugging face ↗

easy run: ollama pull qwen3.8:27b

runs with: ollama · llama.cpp · koboldcpp · lm studio

Alternatives

Mac mini M4 (32GB) — cheapest machine that runs it ($999)

ASUS ExpertCenter Pro ET900N G3 (GB300) — this model on the biggest machine

rtx-5090 — the next step up from this machine's GPU

Buy

NVIDIA (used market) ↗ · Amazon ↗

2026-09-17 · ← full hardware catalog · prices verified at source, may drift