local intelligence

can it run

Can the GTX 970M run qwen3 4B?

TIGHT

3.2 of 3 GB usable — barely over; real with q8 KV cache (halves KV) or shorter context

3GB · Maxwell (laptop) · model file 2.6GB · 4B

~13–23 tok/s

computed roofline: 120.3 GB/s × 0.28–0.50 efficiency window / 2.6GB (Q4_K_M) — bands, never points. Small models (under ~4.5B) run BELOW these bands — kernel-launch overhead keeps them from saturating bandwidth (measured against community medians, 09-06).

Run qwen3 4B on instead

Raspberry Pi 5 + AI Kit (Hailo-8L) — cheapest catalog machine that runs qwen3 4B fully ($110)

every model the GTX 970M can run — full list

explore the full catalog84 machines indexed · live prices · what each one can run
2026-09-17 · ← all models on the GTX 970M · fit = weights + 4k KV vs usable memory