Rent an RTX 4090
NVIDIA GeForce RTX 4090
The fastest consumer GPU with exceptional AI performance.
Specifications
Best for
Renting a RTX 4090
Rent an RTX 4090 by the minute
The RTX 4090 is the card most open models are sized for: 24 GB of GDDR6X, enough for a 27B language model at 4-bit, a current image model at full precision, or a fine-tune of a mid-size model. On ParalonCloud you rent a whole machine with the card in it, not a slice, and pay for the minutes it runs.
A rental opens as Jupyter Lab in your browser, with a terminal, in under a minute. The hourly rate you see when you start is locked for the whole session, billing stops the minute you stop, and there is no contract or minimum spend. Spot machines cost less and can be taken back by their owner with two minutes' notice.
Every RTX 4090 listed here has been benchmarked by us before it is offered. The same card measures differently from one host to the next, from 428 to 922 GB/s of memory bandwidth across the ones we have verified, and each machine's own result is on its page before you pay anything.
What fits
What fits in 24 GB
The question for a 24 GB card is rarely whether a model loads. It is how much memory is left for context once it has.
Measured
Measured on RTX 4090s on this network
These are our own measurements on rented and provider RTX 4090s, each with the article it comes from. Your numbers depend on the host around the card as much as on the card, which is why every machine's own benchmark is on its page.
Steps
How to rent one
- 01
Pick a machine
Open the rent page and filter for RTX 4090. Each machine shows its measured bandwidth, PCIe link, power limit, CPU, RAM, free disk, location and reliability score.
- 02
Start the rental
Choose Jupyter Lab, accept the terms and start. You need credits for one hour to open a session and are charged only for the minutes you use.
- 03
Work, then stop
Use the terminal for pip, conda or apt, run your job, download what you want to keep, and stop. Or drive the whole thing from a script with the rental API.
Limits
Know before you rent
- A rental is an unprivileged container: Docker inside it does not work. Install what you need with pip, conda or apt.
- Access is through Jupyter Lab and its terminal. Other ports are not exposed to the internet.
- Everything on the machine is wiped when the rental ends. Download results before you stop, and checkpoint long jobs.
- A spot machine can be reclaimed with two minutes' notice. Use on demand for work that cannot be interrupted.
FAQ
Questions about the RTX 4090
How much does it cost to rent an RTX 4090?
The current price is shown at the top of this page and comes from the machines available right now, since providers set their own rate. You are billed per minute at the rate shown when you start, and that rate is locked for the session. There is no contract and no minimum beyond having credits for one hour when you open a session.
Can an RTX 4090 run a 27B or 30B class model?
Yes, at 4-bit. We serve a 27B model on single RTX 4090s: the weights take 18.7 GB and about 38,000 tokens of context cache fit in what is left, which is a 32k context. For longer contexts, split the model across two cards.
Is one RTX 4090 enough, or do I need two?
One card gives a 27B model about 30 tokens per second on a single request and 147 in total at 16 concurrent requests. Two cards with the model split across them measured 110 on a single request and 654 at 32, with room for a 131k context. For one user and short prompts, one card is enough. For long context or many users, two.
Do I get SSH or root on the host?
You get Jupyter Lab with a terminal inside an unprivileged container. You can install packages with pip, conda or apt and use the whole GPU, but you cannot run Docker or open other ports.
What is a spot RTX 4090?
A machine whose owner also uses it. It rents at a discount and can be taken back with two minutes' notice, after which the rental stops and billing ends. It suits jobs that checkpoint.
How do I know the card is really an RTX 4090?
Before a machine is listed, a verification job allocates its memory and measures memory bandwidth and compute. The RTX 4090s we have verified measure between 428 and 922 GB/s depending on the host; a card that is not what it claims does not get near that. The result, and when it was last checked, is on the machine's page.
Can I rent one from code?
Yes. The rental API lists available GPUs with prices and starts, inspects and stops rentals with an API key. There are also MCP servers for driving it from an AI coding assistant.