▲ local intelligence

can it run

Can the GTX 980M run LLAMA4 Maverick 400B A17B?

NO

245.8 GB vs 4 GB usable — does not fit

4GB · Maxwell (laptop) · model file 245GB · Maverick 400B A17B

~0.18–0.33 tok/s

computed roofline: 160.4 GB/s × 0.28–0.50 efficiency window / 245.0GB (Q4_K_M) — bands, never points.

Run LLAMA4 Maverick 400B A17B on instead

→ ASUS ExpertCenter Pro ET900N G3 (GB300) — cheapest catalog machine that runs LLAMA4 Maverick 400B A17B fully ($60000)

→ every model the GTX 980M 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 980M · fit = weights + 4k KV vs usable memory