Compute
per instance hour
General Purpose VM4 vCPU / 16 GB RAM
$0.14
Performance VM8 vCPU / 32 GB RAM
$0.32
Bare Metal NodeDedicated host for high-control workloads
$1.85
Pricing
Simple usage pricing for infrastructure and LLM workloads
These rates are reference prices for planning and comparison. Final pricing depends on volume, support model, tenancy, region, and delivery scope.
Storage
per hour / per GB-month equivalent
Block Storage - SSD~$0.16 per GB-month
$0.00022
Block Storage - NVMe~$0.30 per GB-month
$0.00042
Elastic Object Storage~$0.04 per GB-month
$0.00006
LLM Infrastructure & Training
per GPU hour
Training Cluster - A100 80GBPer GPU hour, multi-node available
$4.90
Training Cluster - L40SBalanced option for fine-tuning and medium workloads
$2.80
LLM Platform Base FeePer orchestrated platform hour for control plane and ops
$0.45
Inference
per deployment hour / tokens per second
Shared Inference PodUp to 120 tok/s baseline
$0.85
Dedicated Inference GPUUp to 450 tok/s depending on model size
$2.40
High Throughput Inference Cluster900+ tok/s for larger serving profiles
$6.80
What is included
- Core infrastructure usage for the listed service tier.
- Baseline monitoring and standard operational support windows.
- Clearer planning inputs for compute, storage, and AI workload estimation.
What changes the price
- Reserved capacity, committed volume, and dedicated tenancy.
- Managed service level, compliance requirements, and response-time targets.
- Model size, concurrency, context window, and throughput guarantees for inference.
Need a tailored estimate?
We can turn these reference rates into a workload-based estimate for your compute, storage, or LLM platform needs.