All pricing options
us-ashburn-1 · USD · monthly = 730 hrs
Prices exclude local taxes
| Term / Tier | Hourly | Monthly | Savings |
|---|---|---|---|
On-Demand | $24.40 | $17,812.00 | Baseline |
1-Year Committed Use Discount (CUD) | $16.35 | $11,934.04 | -33% |
3-Year Committed Use Discount (CUD) | $12.20 | $8,906.00 | -50% |
Preemptible / Spot | $12.20 | $8,906.00 | -50% |
You could save $8,906.00/mo on this instance with the right pricing model.
Find my savings Compute
vCPUs i128
Memory i2048 GB
Physical processorAMD EPYC / Intel Xeon
Nested VirtualizationNot supported
Sole TenantNot supported
GPU8 GPU
Lock in this rate across your fleet — typically save 30–40%
Connect your Oracle Cloud tenancy to apply this pricing logic to every running instance automatically.
Networking
Max egress bandwidth1 x 50 Gbps + 8 x 200 Gbps RDMA Gbps
Tier 1 bandwidthN/A
Enhanced networking iYes (gvnic/virtio)
IPv6 supportYes (Dual-stack VPC)
Storage
Max persistent disks16 Volumes
Max disk size32 TB
Local SSDsNo local SSD
BM.GPU4.8 price per GPU-hour
At $24.40 per hour for 8 NVIDIA A100 40GB GPUs, BM_GPU4_8 costs $3.05 per GPU-hour on demand, the rate to use when you compare GPU pricing across providers.
LLM API & GPU Cost Optimization
Running AI or LLM workloads on Oracle Cloud?
Compare OCI compute & GPU instance costs against leading LLM API providers (OpenAI, Anthropic, Google Gemini, DeepSeek, Qwen) and calculate your token pricing side-by-side.
Open LLM Cost Calculator →
Start with AWS 
Start with GCP 
Start with Azure