All providers
Vast.ai
Marketplacespot pricingfree egress
Open marketplace of third-party hosts. Cheapest per GPU-hour, most variable in reliability.
Instance offers
48
32 available right now
Regions
3
Asia Pacific, EU West, US East
Reliability
58/100
Uptime, support and SLA maturity
Cheapest anywhere
8/8
GPU models where this is the market low
Entry price per GPU model
| GPU | Region | vCPU / RAM | Price | Spot | vs market low | Status |
|---|---|---|---|---|---|---|
| A10G 24GB | US East | 8 / 32 GB | $0.278/hr | $0.153/hr | market low | Available |
| RTX 4090 24GB | US East | 8 / 48 GB | $0.341/hr | $0.188/hr | market low | Limited |
| RTX A6000 48GB | US East | 14 / 64 GB | $0.373/hr | $0.205/hr | market low | Sold out |
| L40S 48GB | US East | 16 / 96 GB | $0.539/hr | $0.296/hr | market low | Sold out |
| A100 PCIe 40GB | US East | 12 / 117 GB | $0.678/hr | $0.373/hr | market low | Available |
| A100 SXM 80GB | US East | 12 / 145 GB | $0.841/hr | $0.463/hr | market low | Limited |
| H100 SXM 80GB | US East | 26 / 200 GB | $1.602/hr | $0.881/hr | market low | Available |
| H200 SXM 141GB | US East | 28 / 226 GB | $2.195/hr | $1.207/hr | market low | Available |
Single-GPU instances only, cheapest region per model. Multi-GPU node shapes are in the full marketplace.