GPU Inference Cost Calculator
Price one inference run across every tracked provider
YOUR RUN8 × H100 80 GB/72 h · 3 d/90% utilized576 GPU-hrs billed · 518.4 used
Prices are collected automatically from providers' public pages and can be outdated or wrong. Confirm with the provider before you buy.
72 h · 3 d
1h8h24h3d1wk1mo3mo1yr
90%
Share of rented hours your job actually uses. Idle time is still billed, so it raises your real cost per GPU-hour.
Median of 17$2,183.04Lambda (Lambda Labs)
Most expensive$3,545.28CoreWeave
Max savings$2,577.6073% less than the top price
WHERE 17 PROVIDERS LANDeach dot is one provider's run total
$967.68 · Latitudemedian $2,183.04$3,545.28 · CoreWeave
Real cost per GPU-hr you use$1.87$1.68 list price ÷ 90% utilization
Running 24/7 for a month$9,676.808 GPUs × 720 h at Latitude
Prices are collected automatically from providers' public pages and can be outdated or wrong. Confirm with the provider before you buy.
Every provider for this run
8 × H100 · 72 h · 3 d · 576 GPU-hrs billedRANKPROVIDER$ / GPU-HRRELATIVE COSTRUN TOTALVS CHEAPEST
Prices as of 14 Nov 2026 · read from each provider's own pricing page · billed in full GPU-hours GPU Inference Cost
Price one run across 17 providers
YOUR RUN8 × H100 · 72 h · 90%576 GPU-hrs billed · 518.4 used
Prices are collected automatically from providers' public pages and can be outdated or wrong. Confirm with the provider before you buy.
CHEAPEST FOR THIS RUN$967.68Latitude · $1.68 per GPU-hr Median of 17$2,183.04
Max savings$2,577.60
WHERE 17 PROVIDERS LAND
$967.68$3,545.28
Real cost per used GPU-hr$1.87
Prices are collected automatically from providers' public pages and can be outdated or wrong. Confirm with the provider before you buy.
Every provider
Prices as of 14 Nov 2026 · read from each provider's pricing page · billed in full GPU-hours