NVIDIA L4 rental price
DatacenterLow-power 24 GB inference card. The default cheap accelerator at the hyperscalers.
live · updated 17d ago
VRAM
24 GB GDDR6
Architecture
Ada Lovelace
Segment
Datacenter
Providers quoting
4
best rate
$0.309/GPU-hr at Vast.ai
24h
—
7d
—
30d
—
cheapest 8× node
$7.600/hr
The cheapest NVIDIA L4 you can rent right now is $0.309 per GPU-hour at Vast.ai, against $0.950 at the dearest of the 4 providers quoting it — a 3.1× spread for identical silicon. A full 8-GPU node starts at $7.600/hr at Scaleway. Rates below refresh automatically.
L4 hourly rate history
Daily low $/GPU-hour per provider, cheapest provider first — hover the chart for detail.
Where to rent the NVIDIA L4
Vast.ai
Marketplace
$0.309/GPU-hr
1× L4 · $0.309/hr per node · Arizona, US
- 2× L4 x2$0.619/hr
- 1× L4 x1$0.312/hr
- 1× L4 x1$0.312/hr
AWS
Hyperscaler
$0.805/GPU-hr
1× L4 · $0.805/hr per node · us-east-1
- 1× g6.2xlarge$0.978/hr
- 4× g6.12xlarge$4.602/hr
- 1× g6.4xlarge$1.323/hr
Scaleway
GPU cloud
$0.950/GPU-hr
1× L4 · $0.950/hr per node · pcie · FR, Paris
- 4× L4-4-24G$3.800/hr
- 2× L4-2-24G$1.900/hr
- 8× L4-8-24G$7.600/hr
L4 cluster nodes
Multi-GPU configurations, cheapest per GPU-hour first. For distributed training the interconnect matters as much as the rate — an NVLink or InfiniBand node and the same count of loose PCIe cards are not the same machine.
| Provider | Configuration | Interconnect | $/GPU-hr | Node/hr |
|---|---|---|---|---|
| Vast.ai | 2× L4 x2 · Utah, US | — | $0.309 | $0.619/hr |
| Vast.ai | 4× L4 x4 · Iceland, IS | — | $0.334 | $1.336/hr |
| Scaleway | 4× L4-4-24G · PL, Warsaw | pcie | $0.950 | $3.800/hr |
| Scaleway | 2× L4-2-24G · FR, Paris | pcie | $0.950 | $1.900/hr |
| Scaleway | 8× L4-8-24G · PL, Warsaw | pcie | $0.950 | $7.600/hr |
| AWS | 4× g6.12xlarge · us-east-1 | — | $1.150 | $4.602/hr |
| AWS | 8× g6.48xlarge · us-east-1 | — | $1.669 | $13.35/hr |