▲ local intelligence

can it run

Can the NVIDIA GeForce RTX 4080 run qwen3 8B?

FULL

5.8 GB needed of 16 GB usable — headroom for context

16GB · Ada · model file 5.2GB · 8B

~63–100 tok/s

computed roofline: 716.8 GB/s × 0.46–0.76 efficiency window / 5.2GB (Q4_K_M) — bands, never points.

Run qwen3 8B on instead

→ Tesla M60 (16GB, used) — cheapest catalog machine that runs qwen3 8B fully ($35)

→ rtx-4090 — the next step up from NVIDIA GeForce RTX 4080

→ every model the NVIDIA GeForce RTX 4080 can run — full list

→explore the full catalog91 machines indexed · live prices · what each one can run
2026-10-01 · ← all models on the NVIDIA GeForce RTX 4080 · fit = weights + 4k KV vs usable memory