Pick up to four instance types from one cloud and compare their specs, pricing, and cost efficiency in a single region. Looking for instances by vCPU or memory range instead? Browse by spec.
| GPU-acceleratedp3.2xlarge | GPU-acceleratedp3.8xlarge | |
|---|---|---|
| Compute | ||
| vCPUs | 8 | 32 |
| Memory | 61 GB | 244 GB |
| Memory per vCPU | 7.63 GB | 7.63 GB |
| Architecture | x86_64 | x86_64 |
| Processor | Intel | Intel |
| Family | p3 | p3 |
| Category | GPU-accelerated | GPU-accelerated |
| GPU | ||
| GPUs | 1× V100 | 4× V100 |
| GPU memory | 16 GB | 16 GB |
| $ / GPU / hr | $3.060 | $3.060 |
| Price in us-east-1 | ||
| On-demand | $3.060$2233.80/mo | $12.240$8935.20/mo |
| Spot | $0.379$276.89/mo | $3.762$2746.41/mo |
| Reserved 1-year | — | — |
| Reserved 3-year | — | — |
| Spot savings | 88% | 69% |
| Cost efficiency in us-east-1 | ||
| $ / vCPU / hr | $0.3825 | $0.3825 |
| $ / GB RAM / hr | $0.0502 | $0.0502 |
| Availability | ||
| Regions | 11 | 11 |
| Cheapest region | us-east-1 · $3.060 | us-east-1 · $12.240 |
| Cheapest spot region | us-east-1 · $0.379 | us-east-1 · $3.762 |
| Price snapshot | 2026-04-29 | 2026-04-29 |