Runpod

Runpod H100 Serverless GPU-minute

H100 Serverless workers at $4.55 per active GPU-hour ($0.075833 per minute), billed per second while workers run. Pod H100 rates are lower for steady loads.

Usage calculator

List prices are stored in US dollars. Conversions use the units-per-dollar exchange-rate snapshot last checked 2026-07-31

Published unit price
$0.07583
Estimated cost
$75.833

Project monthly volume

Estimated daily cost
Estimated monthly cost

Cost by usage volume

The table includes intermediate steps such as 25, 50, and 250 units, so you can inspect growth between large round-number bands.

Usage volume Estimated cost
1 $0.07583
10 $0.75833
25 $1.8958
50 $3.7916
100 $7.5833
250 $18.9583
500 $37.9165
1,000 $75.833
2,500 $189.5825
5,000 $379.165
10,000 $758.33
25,000 $1,895.83
50,000 $3,791.65
100,000 $7,583.30
250,000 $18,958.25
500,000 $37,916.50
1,000,000 $75,833.00
5,000,000 $379,165.00
10,000,000 $758,330.00

Cost profile

  • The published rate is $0.07583 per H100 Serverless GPU-minute.
  • One thousand units cost $75.833; 100,000 units cost $7,583.30 before allowances, tiers, or other billing components.
  • A $100 budget covers about 1,318 units at this linear rate.

About Runpod

GPU cloud for pods, serverless inference, and clusters with public per-hour rates. The Runpod H100 Serverless GPU-minute page covers this meter in detail; the Runpod hub gathers the rest of the lineup with a pricing table, list-rate ranges, and calculators side by side.

See Runpod services, APIs, and the pricing table

This calculator applies the published unit rate linearly. Included quotas, tiered prices, minimum commitments, regional charges, and other billing components may change the actual total.

Related services