Vast.ai lists 17 GPU models across 96 configurations, starting at $0.020 per GPU-hour for the V100. It is the cheapest marketplace provider we track for 31 of its 34 GPU-and-pricing-model combinations: V100 (spot), V100 (on-demand), RTX A6000 (spot), RTX PRO 6000 (spot), RTX 4090 (spot), L4 (spot).
| GPU | VRAM | Pricing | Vast.ai $/GPU-hr | Best marketplace | Difference | Cheapest at | |
|---|---|---|---|---|---|---|---|
| V100 | 16 GB | spot | $0.020 | $0.020 | cheapest | — | rent |
| V100 | 16 GB | on-demand | $0.022 | $0.022 | cheapest | — | rent |
| RTX A6000 | 48 GB | spot | $0.053 | $0.053 | cheapest | — | rent |
| RTX PRO 6000 | 96 GB | spot | $0.067 | $0.067 | cheapest | — | rent |
| RTX 4090 | 24 GB | spot | $0.072 | $0.072 | cheapest | — | rent |
| L4 | 24 GB | spot | $0.080 | $0.080 | cheapest | — | rent |
| RTX 4090 | 24 GB | on-demand | $0.120 | $0.120 | cheapest | — | rent |
| A100 SXM | 80 GB | spot | $0.130 | $0.130 | cheapest | — | rent |
| A100 PCIe | 80 GB | spot | $0.133 | $0.133 | cheapest | — | rent |
| H100 SXM | 80 GB | spot | $0.133 | $0.133 | cheapest | — | rent |
| RTX 6000 Ada | 48 GB | spot | $0.133 | $0.133 | cheapest | — | rent |
| L40S | 48 GB | spot | $0.200 | $0.200 | cheapest | — | rent |
| L4 | 24 GB | on-demand | $0.202 | $0.202 | cheapest | — | rent |
| RTX 5090 | 32 GB | spot | $0.267 | $0.267 | cheapest | — | rent |
| RTX A6000 | 48 GB | on-demand | $0.281 | $0.281 | cheapest | — | rent |
| A40 | 48 GB | spot | $0.320 | $0.320 | cheapest | — | rent |
| L40 | 48 GB | spot | $0.320 | $0.320 | cheapest | — | rent |
| RTX 5090 | 32 GB | on-demand | $0.334 | $0.250 | +34% | Salad Cloud | rent |
| L40 | 48 GB | on-demand | $0.335 | $0.330 | +2% | Lium.io | rent |
| A40 | 48 GB | on-demand | $0.336 | $0.336 | cheapest | — | rent |
| H100 PCIe | 80 GB | spot | $0.347 | $0.347 | cheapest | — | rent |
| RTX 6000 Ada | 48 GB | on-demand | $0.388 | $0.388 | cheapest | — | rent |
| H100 NVL | 94 GB | spot | $0.400 | $0.400 | cheapest | — | rent |
| L40S | 48 GB | on-demand | $0.467 | $0.380 | +23% | Lium.io | rent |
| A100 SXM | 80 GB | on-demand | $0.667 | $0.667 | cheapest | — | rent |
| RTX PRO 6000 | 96 GB | on-demand | $0.761 | $0.761 | cheapest | — | rent |
| A100 PCIe | 80 GB | on-demand | $0.870 | $0.870 | cheapest | — | rent |
| H200 | 141 GB | spot | $1.00 | $1.00 | cheapest | — | rent |
| H100 PCIe | 80 GB | on-demand | $1.33 | $1.33 | cheapest | — | rent |
| H100 SXM | 80 GB | on-demand | $1.34 | $1.34 | cheapest | — | rent |
| H100 NVL | 94 GB | on-demand | $2.40 | $2.40 | cheapest | — | rent |
| H200 | 141 GB | on-demand | $2.74 | $2.74 | cheapest | — | rent |
| B200 | 192 GB | spot | $4.38 | $4.38 | cheapest | — | rent |
| B200 | 192 GB | on-demand | $5.38 | $5.38 | cheapest | — | rent |
How much does a GPU cost on Vast.ai?
Vast.ai starts at $0.020 per GPU-hour for the V100 on spot capacity. It lists 17 GPU models in total: A100 PCIe, A100 SXM, A40, B200, H100 NVL, H100 PCIe, H100 SXM, H200, L4, L40.
Is Vast.ai cheaper than other GPU clouds?
For 31 of the 34 GPU and pricing-model combinations it offers, Vast.ai is the cheapest marketplace provider on this table — including V100, V100, RTX A6000, RTX PRO 6000. 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 Vast.ai offer?
17 models: A100 PCIe, A100 SXM, A40, B200, H100 NVL, H100 PCIe, H100 SXM, H200, L4, L40, L40S, RTX 4090, RTX 5090, RTX 6000 Ada, RTX A6000, RTX PRO 6000, V100. Availability changes; the "in stock" flags on each GPU page reflect the last scrape.