Pricing

GPU Container GPU Virtual Machine Serverless Inference Dedicated Inference
USD JPY VND
Pay-per-GPU-hour for the underlying instance hosting your deployed model. Token usage is included — no extra per-token charge. Billing: per-minute · minimum 60 seconds · rounded up

H100 GPU Instances

GPU instance
Specification
Price/hour
1xGPU H100 SXM5
250GB RAM15 cores CPU1TB NVMe Temporary Disk
$3.70
2xGPU H100 SXM5
500GB RAM30 cores CPU2TB NVMe Temporary Disk
$7.40
3xGPU H100 SXM5
750GB RAM45 cores CPU3TB NVMe Temporary Disk
$11.10
4xGPU H100 SXM5
1000GB RAM60 cores CPU4TB NVMe Temporary Disk
$14.80
5xGPU H100 SXM5
1250GB RAM75 cores CPU5TB NVMe Temporary Disk
$18.50
6xGPU H100 SXM5
1500GB RAM90 cores CPU6TB NVMe Temporary Disk
$22.20
7xGPU H100 SXM5
1750GB RAM105 cores CPU7TB NVMe Temporary Disk
$25.90
8xGPU H100 SXM5
2000GB RAM120 cores CPU8TB NVMe Temporary Disk
$29.60

Storage

Storage type
Specification
Location
Price (GB/hour)
Persistent Storage
Suitable for storing data, models, or files that need to be retained even the instance is halted or rebooted. Billed by 1 GB
Southeast Asia, Japan
$0.00023