local intelligence

can it run

Can the NVIDIA A100 80GB SXM run WHISPER large-v3?

FULL

8.2 GB needed of 80 GB usable — headroom for context

80GB · Ampere · model file 6.2GB · large-v3

~150–250 tok/s

computed roofline: 2039 GB/s × 0.46–0.76 efficiency window / 6.2GB (Q4_K_M) — bands, never points.

Run WHISPER large-v3 on instead

Tesla M60 (16GB, used) — cheapest catalog machine that runs WHISPER large-v3 fully ($35)

H100 80GB PCIe (used) — upgrade path machine

every model the NVIDIA A100 80GB SXM can run — full list

explore the full catalog90 machines indexed · live prices · what each one can run
2026-09-24 · ← all models on the NVIDIA A100 80GB SXM · fit = weights + 4k KV vs usable memory