▲ local intelligence

can it run

Can the NVIDIA GeForce GTX 1660 run qwen3 32B?

NO

21.0 GB vs 6 GB usable — does not fit

6GB · Turing · model file 20GB · 32B

~4.6–6.9 tok/s

computed roofline: 192.1 GB/s × 0.48–0.72 efficiency window / 20.0GB (Q4_K_M) — bands, never points.

Run qwen3 32B on instead

→ Tesla V100 32GB (used) — cheapest catalog machine that runs qwen3 32B fully ($700)

→ gtx-1660-super — the next step up from NVIDIA GeForce GTX 1660

→ every model the NVIDIA GeForce GTX 1660 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 GTX 1660 · fit = weights + 4k KV vs usable memory