local intelligence

can it run

Can the Radeon Vega 3 (iGPU) run qwen3 8B?

TIGHT

5.8 of 5.5 GB usable — barely over; real with q8 KV cache (halves KV) or shorter context

~5.5GB usable · GCN5.1 (iGPU) · model file 5.2GB · 8B

~1.8–5.9 tok/s

computed roofline: 38.4 GB/s × 0.25–0.80 efficiency window / 5.2GB (Q4_K_M) — bands, never points.

Run qwen3 8B on instead

Tesla P100 (16GB, used) — cheapest catalog machine that runs qwen3 8B fully ($135)

every model the Radeon Vega 3 (iGPU) can run — full list

explore the full catalog84 machines indexed · live prices · what each one can run
2026-09-17 · ← all models on the Radeon Vega 3 (iGPU) · fit = weights + 4k KV vs usable memory