Skip to content

NVIDIA B200 vs NVIDIA H100

The H100 is 58% cheaper per GPU-hour: $2.50/GPU/hr against $6.00/GPU/hr.

Mean on-demand price per GPU-hour

Line chart with 719 data points. Use the left and right arrow keys to move between data points, Page Up and Page Down to jump by a tenth of the data, Home and End to jump to the first and last, Enter or Space to activate the current point, and Escape to dismiss the readout.

Side by side

NVIDIA B200NVIDIA H100
ArchitectureBlackwellHopper
Memory180 GB HBM3e80 GB HBM3
Cheapest listed$6.00/GPU/hr$2.50/GPU/hr
Cheapest in stock$7.99/GPU/hr$2.50/GPU/hr
Per GB of VRAM$0.0333/hr$0.0313/hr
Reported in stock3.5%24.6%
Providers selling it58
Cheapest you could rentRunPod B200Hyperstack n3-H100x1

What this page doesn't tell you

Which of these is faster for your workload. We price compute. We don't benchmark it. Throughput depends on the model you're running, the precision, the batch size, and the interconnect, and no table of hourly prices substitutes for that.

What it does answer: which one costs less per GPU-hour, which costs less per GB of memory, and how often either was in stock. The cheapest listed price often isn't one you can rent, so that last answer matters most.