▲ local intelligence

can it run

Can the Intel UHD 620 (iGPU) run llava 7B?

OFFLOAD PARTIAL

6.7 GB vs 5.5 GB usable — partial CPU offload, expect large speed loss With a q8 (quantized) KV cache it comes to the edge (5.7GB) — possible with modest context, at some quality cost.

~5.5GB usable · Intel UHD (iGPU) · model file 4.7GB · 7B

~1.2–3.7 tok/s

computed roofline: 38.4 GB/s × 0.15–0.45 efficiency window / 4.7GB (Q4_K_M) — bands, never points.

Run llava 7B on instead

→ Tesla M60 (16GB, used) — cheapest catalog machine that runs llava 7B fully ($35)

→ every model the Intel UHD 620 (iGPU) can run — full list

→explore the full catalog91 machines indexed · live prices · what each one can run
2026-10-01 · ← all models on the Intel UHD 620 (iGPU) · fit = weights + 4k KV vs usable memory