What will your run cost?
Totals at today’s per-provider median listed rates — the same numbers as the provider matrix, multiplied out. Listed ≠ guaranteed capacity, egress and storage are extra, and interruptible tiers can be reclaimed mid-run. The URL updates as you change inputs — share it.
| Provider | $/GPU-hr (median) | 8× · 168h | vs cheapest |
|---|---|---|---|
| Voltage Parkcheapest | $1.99 | $2,675 | — |
| DataCrunch | $3.25 | $4,368 | +$1,693 |
| RunPod | $3.29 | $4,422 | +$1,747 |
| Hyperstack | $3.99 | $5,363 | +$2,688 |
| Lambda | $4.14 | $5,564 | +$2,890 |
| Nebius | $4.50 | $6,048 | +$3,373 |
| Azure | $11.68 | $15,692 | +$13,017 |
Spread on this run: $13,017 between cheapest and most expensive — same silicon.