local intelligence

can it run

Can the NVIDIA GTX 1660 Super run llama3.3 70B?

NO

44.2 GB vs 6 GB usable — does not fit

6GB · Turing TU116 · model file 43GB · 70B

~3.8–5.6 tok/s

computed roofline: 336 GB/s × 0.48–0.72 efficiency window / 43.0GB (Q4_K_M) — bands, never points.

Run llama3.3 70B on instead

NIMO AI Mini PC (128GB) — cheapest catalog machine that runs llama3.3 70B fully ($1999)

rtx-5060-ti-16 — the next step up from NVIDIA GTX 1660 Super

every model the NVIDIA GTX 1660 Super can run — full list

explore the full catalog84 machines indexed · live prices · what each one can run
2026-09-17 · ← all models on the NVIDIA GTX 1660 Super · fit = weights + 4k KV vs usable memory