Runpod lists 16 GPU models across 32 configurations, starting at $0.440 per GPU-hour for the A40. It is the cheapest dedicated provider we track for 3 of its 32 GPU-and-pricing-model combinations: A40 (on-demand), A40 (spot), H100 PCIe (spot).
| GPU | VRAM | Pricing | Runpod $/GPU-hr | Best dedicated | Difference | Cheapest at | |
|---|---|---|---|---|---|---|---|
| A40 | 48 GB | on-demand | $0.440 | $0.440 | cheapest | — | rent |
| A40 | 48 GB | spot | $0.440 | $0.440 | cheapest | — | rent |
| L4 | 24 GB | on-demand | $0.490 | $0.440 | +11% | Jarvis Labs | rent |
| L4 | 24 GB | spot | $0.490 | $0.056 | +783% | Google Cloud | rent |
| RTX A6000 | 48 GB | on-demand | $0.530 | $0.350 | +51% | Thunder Compute | rent |
| RTX A6000 | 48 GB | spot | $0.530 | $0.305 | +74% | Verda | rent |
| RTX 4090 | 24 GB | on-demand | $0.740 | $0.340 | +118% | GPU.ai | rent |
| RTX 4090 | 24 GB | spot | $0.740 | $0.340 | +118% | Novita | rent |
| RTX 6000 Ada | 48 GB | on-demand | $0.840 | $0.750 | +12% | Prime Intellect | rent |
| RTX 6000 Ada | 48 GB | spot | $0.840 | $0.520 | +62% | Verda | rent |
| L40S | 48 GB | on-demand | $0.990 | $0.550 | +80% | Novita | rent |
| L40S | 48 GB | spot | $0.990 | $0.423 | +134% | AWS | rent |
| RTX 5090 | 32 GB | on-demand | $0.990 | $0.730 | +36% | Novita | rent |
| RTX 5090 | 32 GB | spot | $0.990 | $0.370 | +168% | Novita | rent |
| A100 PCIe | 80 GB | on-demand | $1.39 | $1.20 | +16% | Prime Intellect | rent |
| A100 PCIe | 80 GB | spot | $1.39 | $0.679 | +105% | Azure | rent |
| A100 SXM | 80 GB | on-demand | $1.59 | $1.09 | +46% | Thunder Compute | rent |
| A100 SXM | 80 GB | spot | $1.59 | $0.410 | +287% | Google Cloud | rent |
| RTX PRO 6000 | 96 GB | on-demand | $1.89 | $1.35 | +40% | Prime Intellect | rent |
| RTX PRO 6000 | 96 GB | spot | $1.89 | $0.945 | +100% | Verda | rent |
| H100 PCIe | 80 GB | on-demand | $2.89 | $2.35 | +23% | Prime Intellect | rent |
| H100 PCIe | 80 GB | spot | $2.89 | $2.89 | cheapest | — | rent |
| H100 NVL | 94 GB | on-demand | $3.19 | $3.11 | +3% | Massed Compute | rent |
| H100 NVL | 94 GB | spot | $3.19 | $1.29 | +147% | Azure | rent |
| H100 SXM | 80 GB | on-demand | $3.29 | $1.79 | +84% | Cudo Compute | rent |
| H100 SXM | 80 GB | spot | $3.29 | $0.970 | +239% | Google Cloud | rent |
| H200 | 141 GB | on-demand | $3.79 | $3.03 | +25% | Seeweb | rent |
| H200 | 141 GB | spot | $3.79 | $1.85 | +105% | AWS | rent |
| B200 | 192 GB | on-demand | $6.79 | $4.90 | +39% | GPU.ai | rent |
| B200 | 192 GB | spot | $6.79 | $1.81 | +275% | Google Cloud | rent |
| B300 | 288 GB | on-demand | $7.89 | $6.60 | +20% | Massed Compute | rent |
| B300 | 288 GB | spot | $7.89 | $3.69 | +114% | AWS | rent |
How much does a GPU cost on Runpod?
Runpod starts at $0.440 per GPU-hour for the A40. It lists 16 GPU models in total: A100 PCIe, A100 SXM, A40, B200, B300, H100 NVL, H100 PCIe, H100 SXM, H200, L4.
Is Runpod cheaper than other GPU clouds?
For 3 of the 32 GPU and pricing-model combinations it offers, Runpod is the cheapest dedicated provider on this table — including A40, A40, H100 PCIe. For the rest, the table above names the provider that beats it and by how much. Comparisons are like-for-like on GPU, pricing model and capacity tier.
Which GPUs does Runpod offer?
16 models: A100 PCIe, A100 SXM, A40, B200, B300, H100 NVL, H100 PCIe, H100 SXM, H200, L4, L40S, RTX 4090, RTX 5090, RTX 6000 Ada, RTX A6000, RTX PRO 6000. Availability changes; the "in stock" flags on each GPU page reflect the last scrape.