local intelligence

can it run

Can the NVIDIA GeForce RTX 3070 Ti run qwen3 14B?

OFFLOAD PARTIAL

9.9 GB vs 8 GB usable — partial CPU offload, expect large speed loss

8GB · Ampere · model file 9.3GB · 14B

~30–50 tok/s

computed roofline: 608.3 GB/s × 0.46–0.76 efficiency window / 9.3GB (Q4_K_M) — bands, never points.

Run qwen3 14B on instead

Tesla P100 (16GB, used) — cheapest catalog machine that runs qwen3 14B fully ($135)

rtx-3080-10gb — the next step up from NVIDIA GeForce RTX 3070 Ti

every model the NVIDIA GeForce RTX 3070 Ti 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 GeForce RTX 3070 Ti · fit = weights + 4k KV vs usable memory