NVIDIA RTX 4090 Pricing
Eli Mernit
An NVIDIA RTX 4090 costs $0.20 to $0.44 per GPU-hour to rent on demand from the providers below, against a market median of $0.47 across 14 providers. Beam's RTX 4090 machine is $0.44 an hour with vCPU, RAM, and NVMe included. Beam's serverless 4090 bills $0.69 an hour for the GPU, plus CPU and RAM, and scales to zero when idle.
Key takeaways
- The $0.20 floor is the cheapest single offer on Vast.ai's marketplace. The median Vast.ai offer is $0.48, and nine in ten cost at least $0.40.
- Salad High priority ($0.33) and RunPod Community Cloud ($0.34) are the cheapest fixed rates, both below Beam's $0.44.
- Beam's machine is $0.03 under the market median, and its serverless option pays off for spiky traffic.
RTX 4090 cloud prices per hour, by provider
| Provider | SKU | On-demand $/GPU-hr | Notes |
|---|---|---|---|
| Vast.ai | RTX 4090 (marketplace) | $0.20 | Cheapest rentable offer; median offer $0.48; spot from $0.16 per getdeploying; marketplace data from Vast.ai |
| Salad | RTX 4090 (24 GB), High priority | $0.33 | Consumer-PC network; vCPU and RAM included; interruptible tiers from $0.16; in stock per getdeploying |
| RunPod | RTX 4090 (Community Cloud) | $0.34 | Secure Cloud $0.74; both in stock per getdeploying |
| Beam | RTX 4090 PCIe | $0.44 | On-demand machine; vCPU, RAM, and NVMe included; serverless GPU $0.69/hr on every plan, plus CPU and RAM |
| Market median | — | $0.47 | Across 14 providers with an on-demand price (getdeploying) |
How these prices were collected
Each figure is an on-demand, per-GPU list price for a single-GPU shape where one is offered. Spot and interruptible rates appear only in the notes. Marketplace data from Vast.ai.
- Vast.ai's GPU pricing page has the floor and median offers.
- Salad's pricing page lists each priority tier.
- RunPod's pricing page has both cloud tiers.
- Beam's pricing page has the on-demand and serverless rates.
- getdeploying's RTX 4090 tracker supplies the median.
RTX 4090 marketplace floors vs fixed-rate listings
The cheapest RTX 4090s come from marketplaces and consumer hardware. On Vast.ai, hosts set their own prices, and only one offer in ten costs $0.40 or less. Salad runs on consumer PCs: its High priority tier is never preempted by other workloads, but a node can still disconnect (Salad reallocates the workload automatically), and cheaper tiers can be preempted. That suits retry-safe inference more than long training runs. Fixed-rate listings run from $0.33 (Salad High priority) to $0.74 (RunPod Secure Cloud).
Where to rent an RTX 4090 for less
Three listings undercut Beam's $0.44: Vast.ai from $0.20, Salad at $0.33, and RunPod Community Cloud at $0.34. Salad and RunPod include vCPU and RAM, and getdeploying shows both in stock. If the lowest hourly rate matters most, start there.
Beam's price buys a fixed rate with NVMe included and a serverless option. Serverless has the higher GPU rate ($0.69 against $0.44) plus CPU and RAM charges, so it only saves money when a GPU would otherwise sit idle for long stretches.
What this means for your workload
The RTX 4090 handles inference and image generation for models that fit in 24 GB. Retry-safe requests suit marketplace and interruptible capacity. Bursty traffic suits serverless, as long as you plan for cold starts on serverless GPUs after idle stretches. Steady, round-the-clock serving belongs on a fixed-rate machine. If you need more memory, the RTX 5090 has 32 GB at a $0.64 median. Weighing a purchase instead? Our RTX 4090 price guide covers what the card costs to buy.
FAQ
How much does an RTX 4090 cost per hour?
It costs $0.20 to $0.44 per GPU-hour on demand across the providers above, with a market median of $0.47 across 14 providers. Beam's machine is $0.44 an hour, and its serverless 4090 is $0.69 an hour for the GPU on every plan.
What's the cheapest way to rent an RTX 4090?
Vast.ai's marketplace starts at $0.20, but its median offer is $0.48. If your job can be preempted, Salad's interruptible tiers start at $0.16. The cheapest fixed rates here are Salad High priority ($0.33) and RunPod Community Cloud ($0.34).
RTX 4090 vs RTX 5090: which should I rent?
Rent the 4090 if your model fits in 24 GB; its median is $0.47 against $0.64 for the 5090. Pay for the 5090 when you need 32 GB. On Beam, the machines cost $0.44 and $0.72 an hour.
