How Much VRAM Do I Need for LLM Training and Inference?
A practical engineer's guide to VRAM for LLM training, fine-tuning, and inference. Learn how to estimate VRAM for 7B and 70B...
2026-10-25 Read guideCheapest single GPU right now
$0.1600
Per GPU per hour, on-demand, across all tracked models.
- GPU listings
- 17
- Spot capacity
- Yes
- NVLink
- Yes
- Price data last updated
- 2026-10-26
30-day price history
Each point is a daily snapshot we collected ourselves. Solid line = on-demand, dashed = spot.
On-demand
Spot / interruptible
Source: our own daily collection snapshots
Live GPU price comparison
Filter by provider, GPU or region — filters run locally in your browser.
| Provider | GPU model | VRAM | GPUs | vCPU | Region | On-demand /hr | Spot /hr | Actions |
|---|---|---|---|---|---|---|---|---|
| RunPod | RTX A5000 | 24 GB | 1 | — | Global (secure + community cloud) | $0.1600 Best price | — | Rent this GPU |
| RunPod | RTX A4000 | 16 GB | 1 | — | Global (secure + community cloud) | $0.1700 Best price | — | Rent this GPU |
| RunPod | V100 | 16 GB | 1 | — | Global (secure + community cloud) | $0.1900 Best price | — | Rent this GPU |
| RunPod | RTX 3090 | 24 GB | 1 | — | Global (secure + community cloud) | $0.2200 Best price | — | Rent this GPU |
| RunPod | RTX A6000 | 48 GB | 1 | — | Global (secure + community cloud) | $0.3300 Best price | — | Rent this GPU |
| RunPod | RTX 4090 | 24 GB | 1 | — | Global (secure + community cloud) | $0.3400 Best price | — | Rent this GPU |
| RunPod | A40 | 48 GB | 1 | — | Global (secure + community cloud) | $0.3500 Best price | — | Rent this GPU |
| RunPod | L4 | 24 GB | 1 | — | Global (secure + community cloud) | $0.4400 Best price | — | Rent this GPU |
| RunPod | H200 SXM | 141 GB | 1 | — | Global (secure + community cloud) | $3.5900 Best price | — | Rent this GPU |
| RunPod | RTX 5090 | 32 GB | 1 | — | Global (secure + community cloud) | $0.6900 Best price | — | Rent this GPU |
| RunPod | RTX 6000 Ada | 48 GB | 1 | — | Global (secure + community cloud) | $0.7400 Best price | — | Rent this GPU |
| RunPod | L40S | 48 GB | 1 | — | Global (secure + community cloud) | $0.7900 Best price | — | Rent this GPU |
| RunPod | A100 SXM | 80 GB | 1 | — | Global (secure + community cloud) | $1.3900 Best price | — | Rent this GPU |
| RunPod | A100 PCIe | 80 GB | 1 | — | Global (secure + community cloud) | $1.1900 Best price | — | Rent this GPU |
| RunPod | H100 PCIe | 80 GB | 1 | — | Global (secure + community cloud) | $1.9900 Best price | — | Rent this GPU |
| RunPod | H100 SXM | 80 GB | 1 | — | Global (secure + community cloud) | $2.6900 Best price | — | Rent this GPU |
| RunPod | B200 | 180 GB | 1 | — | Global (secure + community cloud) | $5.9800 Best price | — | Rent this GPU |
No offers match these filters. Relax the price limit or clear the search.
Frequently asked questions
Everything about renting cloud GPUs, affiliate tracking and how we collect pricing data.
How is this provider's GPU pricing structured?
Almost all cloud GPU providers bill by the second or the hour for the compute, then separately for persistent storage and outbound data transfer. The figures on this page are the compute rate for one GPU, normalised so you can compare different instance shapes without doing the arithmetic yourself.
Can I get interruptible or spot capacity here?
That depends on the provider and is shown in the availability column. Marketplace-style platforms usually offer a bid-based interruptible tier, while dedicated clouds tend to sell only on-demand capacity. If spot is available it is almost always cheaper, provided your workload can checkpoint and restart.
Is there a minimum contract or commitment?
For on-demand and spot capacity there generally is not: you pay per hour or per second and can stop at any time. Reserved contracts exist at some providers and can cut the effective rate, but they are not required for the prices shown on this page.
How are storage and egress charged?
Persistent volumes are billed per GB per month, and outbound transfer is charged per GB beyond whatever allowance the provider includes. Those costs are additive to the GPU hour and can dominate the bill on data-heavy pipelines, so budget for them explicitly.