AMD · CDNA 3
Instinct MI300X 192 GB rental prices and specifications
An AMD OAM accelerator with 192 GB of HBM3 and ROCm support, commonly deployed in eight-GPU systems for large-model inference and training.
Profile checked · Primary source ↗
- Memory
- 192 GB
- Bandwidth
- 5300 GB/s
- Current offers
- 0
- FP16 dense
- 1307.4 TFLOPS
- BF16 dense
- 1307.4 TFLOPS
- FP8 dense
- 2614.9 TFLOPS
- Interconnect
- Seven 128 GB/s Infinity Fabric links (896 GB/s aggregate) per MI300X OAM; PCIe Gen 5 x16 host links. Provider topology requires separate evidence.
Compare Instinct MI300X 192 GB cloud rentals
Use the exact 192 GB GPU profile when comparing rental quotes. Tables separate currencies, billing periods and rental terms; the required GPU count determines the whole-instance compute cost. Storage, transfer, taxes and other charges can be additional.
Source checks show when each price was actually observed. Historical quotes remain available separately and do not enter current price comparisons. Explore the sourced GPU rental price index or check a rental against your model requirements.
Manufacturer theoretical peak matrix throughput per OAM; dense values, not measured model throughput. · Specification source 1 ↗ · Specification source 2 ↗
Recently checked listings
Current rental offers
| GPU / provider | Instance price / offer | Terms / location |
|---|---|---|
| No offers match these filters. Try another billing period or currency. | ||
Previously observed offers
6 retained quotes with an old check, a different allocation, or unavailable status. These are excluded from current prices and capacity counts.
Hide previously observed offers
USD · Hourly · On-demand
| GPU / provider | Instance price / offer | Terms / location |
|---|---|---|
| Instinct MI300X 192 GB Microsoft Azure Standard_ND96isr_MI300X_v5 192 GB per GPU · 8 GPUs per instance | Previously observed quote USD 59.04/hr per instanceUSD 7.38 /hr per GPU Browse Microsoft Azure Details & price history Native rate: USD 59.041 / hour per instance Saved pricing evidence → Pricing evidence (API response) Readable pricing page Stale — recheck | HourlyOn-demand Minimum term unverified Stock not verified · United States · Central US |
| Instinct MI300X 192 GB Hot Aisle Large 8x MI300X bare metal 192 GB per GPU · 8 GPUs per instance | Previously observed quote USD 27.12/hr per instanceUSD 3.39 /hr per GPU Browse Hot Aisle Details & price history Pricing evidence Stale — recheck |
HourlyOn-demand
Minimum term unverified
Stock not verified · United States · Michigan
Billing & region detailsHourly advertised bare-metal rate; minimum term and billing increments not verified. |
| Instinct MI300X 192 GB DigitalOcean GPU Droplet AMD MI300X 1x On-demand 192 GB per GPU · 1 GPU per instance | Previously observed quote USD 2.59/hr per instanceUSD 2.59 /hr per GPU Browse DigitalOcean Details & price history Pricing evidence Stale — recheck |
HourlyOn-demand
Minimum term unverified
Stock not verified · Location not specified
Billing & region detailsPer-second usage billing; minimum billable duration needs confirmation. Billing continues while powered off until the Droplet is destroyed. |
| Instinct MI300X 192 GB DigitalOcean GPU Droplet AMD MI300X 8x On-demand 192 GB per GPU · 8 GPUs per instance | Previously observed quote USD 20.72/hr per instanceUSD 2.59 /hr per GPU Browse DigitalOcean Details & price history Pricing evidence Stale — recheck |
HourlyOn-demand
Minimum term unverified
Stock not verified · Location not specified
Billing & region detailsPer-second usage billing; minimum billable duration needs confirmation. Billing continues while powered off until the Droplet is destroyed. |
| Instinct MI300X 192 GB Crusoe Cloud mi300x-192gb-ib.8x 192 GB per GPU · 8 GPUs per instance | Previously observed quote USD 27.60/hr per instanceUSD 3.45 /hr per GPU Browse Crusoe Cloud Details & price history Pricing evidence Stale — recheck |
HourlyOn-demand
No monthly commitment
Stock not verified · United States · us-east1-a
Billing & region detailsNo minimum monthly commitment. Billed per second only while Running, including lifecycle scripts. Compute stops billing when stopped; the OS disk remains billable until the VM is deleted. |
Prices and availability can change. Confirm the current offer with the provider. Taxes and extra services are additional.
USD · Hourly · Spot / interruptible
| GPU / provider | Instance price / offer | Terms / location |
|---|---|---|
| Instinct MI300X 192 GB Microsoft Azure Standard_ND96isr_MI300X_v5 Spot 192 GB per GPU · 8 GPUs per instance | Previously observed quote USD 10.91/hr per instanceUSD 1.36 /hr per GPU Browse Microsoft Azure Details & price history Native rate: USD 10.910777 / hour per instance Saved pricing evidence → Pricing evidence (API response) Readable pricing page Stale — recheck |
HourlySpot / interruptible
Minimum term unverified
Stock not verified · United States · Central US
Billing & region detailsAzure Spot capacity is interruptible and has no availability SLA. |
Prices and availability can change. Confirm the current offer with the provider. Taxes and extra services are additional.
Measured performance
Inference benchmarks
| Model | GPU setup | Engine | Quant. | Workload | Reported tok/s | Evidence | Max-load capacity scenarioUSD / million output tokens |
|---|---|---|---|---|---|---|---|
| No benchmark records have been supplied yet. | |||||||