Cloud GPU per GB of VRAM: cheapest among tracked providers
Hourly rate divided by GPU memory. Useful when you need a model to fit and are deciding which card to rent.
| # | GPU | Vendor | Type | Rate | $ per GB VRAM-hour | Vendor quote |
|---|---|---|---|---|---|---|
| 1 | RTX A6000 48 GB | RunPod | Community Cloud | $0.33/hr | $0.0069 | “RTX A6000 $0.33/hr” |
| 2 | RTX A6000 48 GB | Thunder Compute | on-demand | $0.35/hr | $0.0073 | “RTX A6000 48GB $0.35/hr” |
| 3 | RTX 3090 24 GB | RunPod | Community Cloud | $0.22/hr | $0.0092 | “RTX 3090 $0.22/hr” |
| 4 | RTX A5000 24 GB | RunPod | Secure Cloud | $0.27/hr | $0.0113 | “RTX A5000 24GB $0.27/hr” |
| 5 | RTX A6000 48 GB | RunPod | Secure Cloud | $0.59/hr | $0.0123 | “RTX A6000 48GB $0.59/hr” |
| 6 | A100 80 GB | Thunder Compute | on-demand | $1.09/hr | $0.0136 | “A100 80GB $1.09/hr” |
| 7 | RTX 4090 24 GB | RunPod | Community Cloud | $0.34/hr | $0.0142 | “RTX 4090 $0.34/hr” |
| 8 | A100 PCIe 80 GB | RunPod | Community Cloud | $1.19/hr | $0.0149 | “A100 PCIe $1.19/hr” |
| 9 | L40S 48 GB | RunPod | Community Cloud | $0.79/hr | $0.0165 | “L40S $0.79/hr” |
| 10 | L40 48 GB | Thunder Compute | on-demand | $0.79/hr | $0.0165 | “L40 48GB $0.79/hr” |
| 11 | L40 48 GB | RunPod | Secure Cloud | $0.82/hr | $0.0171 | “L40 48GB $0.82/hr” |
| 12 | A100 SXM 80 GB | RunPod | Community Cloud | $1.39/hr | $0.0174 | “A100 SXM $1.39/hr” |
| 13 | RTX 3090 24 GB | RunPod | Secure Cloud | $0.5/hr | $0.0208 | “RTX 3090 24GB $0.50/hr” |
| 14 | RTX 5090 32 GB | RunPod | Community Cloud | $0.69/hr | $0.0216 | “RTX 5090 $0.69/hr” |
| 15 | A100 PCIe 80 GB | RunPod | Secure Cloud | $1.79/hr | $0.0224 | “A100 PCIe 80GB $1.79/hr” |
| 16 | A100 SXM 80 GB | RunPod | Secure Cloud | $1.79/hr | $0.0224 | “A100 SXM 80GB $1.79/hr” |
| 17 | L40S 48 GB | RunPod | Secure Cloud | $1.09/hr | $0.0227 | “L40S 48GB $1.09/hr” |
| 18 | H100 PCIe 80 GB | RunPod | Community Cloud | $1.99/hr | $0.0249 | “H100 PCIe $1.99/hr” |
| 19 | H200 141 GB | RunPod | Community Cloud | $3.59/hr | $0.0255 | “H200 $3.59/hr” |
| 20 | A100 80 GB | Modal | Serverless (per second) | $2.4984/hr | $0.0312 | “Nvidia A100, 80 GB $0.000694/sec” |
| 21 | H200 SXM 141 GB | Modal | Serverless (per second) | $4.5396/hr | $0.0322 | “Nvidia H200 SXM $0.001261/sec” |
| 22 | H100 SXM 80 GB | RunPod | Community Cloud | $2.69/hr | $0.0336 | “H100 SXM $2.69/hr” |
| 23 | H100 80 GB | 99Stack Cloud | on-demand | $2.8767/hr | $0.0360 | “NVIDIA H100 (12 vCPU/240 GB RAM) $2.8767/hr” |
| 24 | H100 PCIe 80 GB | RunPod | Secure Cloud | $2.89/hr | $0.0361 | “H100 PCIe 80GB $2.89/hr” |
| 25 | RTX 4090 24 GB | RunPod | Secure Cloud | $0.89/hr | $0.0371 | “RTX 4090 24GB $0.89/hr” |
| 26 | RTX 5090 32 GB | RunPod | Secure Cloud | $1.19/hr | $0.0372 | “RTX 5090 32GB $1.19/hr” |
| 27 | H200 141 GB | RunPod | Secure Cloud | $5.29/hr | $0.0375 | “H200 141GB $5.29/hr” |
| 28 | H100 SXM 80 GB | Thunder Compute | on-demand | $3.2/hr | $0.0400 | “H100 SXM 80GB $3.20/hr” |
| 29 | L40S 48 GB | Modal | Serverless (per second) | $1.9512/hr | $0.0406 | “Nvidia L40S $0.000542/sec” |
| 30 | H100 SXM 80 GB | Modal | Serverless (per second) | $3.9492/hr | $0.0494 | “Nvidia H100 SXM5 $0.001097/sec” |
| 31 | H100 SXM 80 GB | RunPod | Secure Cloud | $3.99/hr | $0.0499 | “H100 SXM 80GB $3.99/hr” |
Read before comparing
Rankings are the cheapest among tracked providers on our check date, not across the whole market. Vendors we don't track may cost less.
GPU-hours vs wall-clock time
Prices are per GPU-hour: one GPU for one hour. Four GPUs for 10 wall-clock hours is 40 GPU-hours. Per-second billing (serverless) charges only while your code runs; hourly pods bill from start to stop, idle time included; monthly servers bill a flat price whatever you use. A faster GPU can finish the same job in fewer GPU-hours, so the lowest hourly rate isn't always the lowest cost per job.
Hardware differences
The same GPU name can mean different hardware. SXM modules link GPUs over NVLink and run at higher power limits than PCIe cards, so an H100 SXM is usually faster than an H100 PCIe, especially for multi-GPU training. Interconnect between GPUs and nodes (NVLink, InfiniBand, plain Ethernet) is rarely listed on pricing pages; treat it as unknown unless the vendor states it. VRAM decides which models fit at all. Hardware as listed: A100 80 GB (not stated); A100 PCIe 80 GB (PCIe); A100 SXM 80 GB (SXM); H100 80 GB (not stated); H100 PCIe 80 GB (PCIe); H100 SXM 80 GB (SXM); H200 141 GB (not stated); H200 SXM 141 GB (SXM); L40 48 GB (not stated); L40S 48 GB (not stated); RTX 3090 24 GB (not stated); RTX 4090 24 GB (not stated); RTX 5090 32 GB (not stated); RTX A5000 24 GB (not stated); RTX A6000 48 GB (not stated).
Egress and storage extras
Totals on this page exclude these. Each cell cites the vendor page it came from; unknown means the vendor doesn't publish it there.
| Vendor | Egress | Storage | Interruption / billing terms |
|---|---|---|---|
| 99Stack Cloud | unknown: Not published on the product page. (checked 2026-10-27) | unknown: Not published on the product page. (checked 2026-10-27) | Hourly GPU VPS ($2,100/mo shown as the monthly price). (source, 2026-10-27) |
| Modal | Network egress $0.04/GiB; Starter plan includes 1 TiB/month free. (source, 2026-10-27) | Volumes $0.09/GiB/mo (includes 1 TiB/mo free). (source, 2026-10-27) | Functions are preemptible by default; non-preemptible execution is 3x the base price. CPU and memory are billed separately. (source, 2026-10-27) |
| RunPod | unknown: Not published on the pricing page. (checked 2026-10-27) | Container disk $0.10/GB/mo; volume disk $0.10/GB/mo running, $0.20/GB/mo idle; network storage $0.07/GB/mo under 1 TB, $0.05/GB/mo over 1 TB. (source, 2026-10-27) | Listed rates are on-demand Pod prices (Secure Cloud and Community Cloud). (source, 2026-10-27) |
| Thunder Compute | No egress fees ($0.00/GB). (source, 2026-10-27) | Instance disk: first 100 GB per GPU included, then $0.219/GB/mo while running; snapshots $0.05/GB/mo. (source, 2026-10-27) | Instances and sandboxes are never preempted. Instances (including extra vCPUs and disk) are billed per minute while running. (source, 2026-10-27) |
Interruptible vs on-demand
Interruptible (preemptible or spot) capacity can be stopped by the vendor mid-job and is usually cheaper; on-demand runs until you stop it. Interruptible (preemptible) rates listed here: Modal A100, Modal H100 SXM, Modal H200 SXM, Modal L40S. All other rates are on-demand: the vendor doesn't publish preemption for them.
FAQ
Which cloud GPU is cheapest per GB of VRAM among tracked providers?
On 2026-10-27: RunPod RTX A6000 48 GB (Community Cloud) at $0.0069 per GB-hour ($0.33/hr).
Which 80 GB+ GPU is cheapest per GB among tracked providers?
Thunder Compute A100 80 GB (on-demand) at $0.0136 per GB-hour.
Why rank by VRAM?
Model size usually decides which GPU you can use at all. Price per GB of VRAM shows the lowest-cost way among tracked providers to fit a model in memory; it ignores speed, so a faster GPU can still finish a job for less.
Which rates are included?
Only hourly rates checked on 2026-10-27 on the vendor's own pricing page. Monthly dedicated servers are excluded.
More
- All verified GPU rental prices
- H100: cheapest among tracked providers
- H200: cheapest among tracked providers
- A100: cheapest among tracked providers
- L40S: cheapest among tracked providers
- RTX 4090: cheapest among tracked providers
- RTX 5090: cheapest among tracked providers
- RTX A6000: cheapest among tracked providers
- RunPod vs Thunder Compute vs Modal (H100)
- RunPod Community vs Secure Cloud
- GPU cost calculator
- GPU & compute deals
Using an AI agent? Add Lantern Works to your agent (free MCP server, read-only).