NVIDIA · Architecture pending

H100 PCIe 80 GB

VRAM as advertised; benchmark your workload separately.

Profile checked · Primary source ↗

Memory
80 GB
Bandwidth
—
Offers
1

Available capacity

Observed rental offers

GPU / providerBilling & commitmentPer GPUInstance totalStock & regionChecked / evidence
H100 PCIe 80 GBRunpodPublic Pods listing — tier unverified · 80 GB per GPU HourlyOn-demandNo monthly commitmentOn-demand Pods; confirm cloud tier and region in console. Reference onlyUSD 2.89/hr per GPU Minimum GPU count unverified Stock not verifiedRegion not specified Source ↗Details & history →

Measured performance

Inference benchmarks

Model GPU Engine Quant. Tokens / sec TTFT
No benchmark records have been supplied yet.