L40S GPUs
Dedicated L40S instances rented per hour: SSH-rentable containers and dedicated LLM deployment pods. Prices are synced from provider APIs.
L40S GPUs rental offers
| Configuration | GPU | $/hour | Status | Kind | Provider | Last sync |
|---|---|---|---|---|---|---|
14
1xL40S-48GB
|
L40S-48GB | $1.57 | Available | LLM deploy | DigitalOcean | |
|
84
8xL40S-48GB
|
L40S-48GB | $18 | Available | LLM deploy | CoreWeave |
Availability and prices reflect the last sync (3h ago). Hourly billing, no token charges on dedicated hardware.
Frequently asked questions
How is dedicated GPU rental billed?
Per hour of instance uptime, in USD. Unlike serverless inference there are no token charges — you rent the whole card or pod and can run any workload on it, 24/7 or spot.
What is the difference between an LLM deployment pod and an SSH container?
A pod comes with a model-serving stack preinstalled, so you can deploy and scale an LLM in minutes. An SSH container is a plain machine you fully control: install anything, train, fine-tune or serve multiple models.
Where do these prices come from?
Straight from each provider's public infrastructure API during the nightly sync. The "last checked" timestamp on every row shows when that offer was last verified.