All workspaces
GPU Cloud Comparison
Compare GPU cloud providers on hourly price, VRAM, egress and storage cost, billing granularity, and cold-start latency.
All GPU instances
Filters
82 of 82
| 1 | RTX 2060Vast.ai | 0.02 USD/hr | 6 GB | 99.6 % | — | — | — | |
| 2 | GTX 1060Vast.ai | 0.02 USD/hr | 6 GB | 99.9 % | — | — | — | |
| 2 | B200Vast.ai | 3.99 USD/hr | 179 GB | 99.9 % | — | — | — | |
| 3 | GTX 1050 TiVast.ai | 0.01 USD/hr | 4 GB | 94.9 % | — | — | — | |
| 4 | Tesla P4Vast.ai | 0.02 USD/hr | 8 GB | 99.6 % | — | — | — | |
| 5 | RTX 3080Vast.ai | 0.03 USD/hr | 10 GB | 100 % | — | — | — | |
| 6 | H200 NVLVast.ai | 2.6 USD/hr | 140 GB | 99.8 % | — | — | — | |
| 7 | H200Vast.ai | 3.3 USD/hr | 140 GB | 100 % | — | — | — | |
| 8 | L4Vast.ai | 0.03 USD/hr | 22 GB | 99.8 % | — | — | — | |
| 9 | RTX PRO 6000 WSVast.ai | 0.93 USD/hr | 96 GB | 99.9 % | — | — | — | |
| 10 | GTX TITAN XVast.ai | 0.03 USD/hr | 12 GB | 99.4 % | — | — | — | |
| 11 | RTX 2080 TiVast.ai | 0.04 USD/hr | 11 GB | 99.9 % | — | — | — | |
| 12 | RTX PRO 6000 SVast.ai | 1.07 USD/hr | 96 GB | 99.9 % | — | — | — | |
| 13 | RTX 3070Vast.ai | 0.04 USD/hr | 8 GB | 99.9 % | — | — | — | |
| 13 | H100 SXMVast.ai | 0.28 USD/hr | 80 GB | 99.9 % | — | — | — | |
| 13 | RTX A2000Vast.ai | 0.04 USD/hr | 6 GB | 99.9 % | — | — | — | |
| 14 | GTX 1660 SVast.ai | 0.04 USD/hr | 6 GB | 99.6 % | — | — | — | |
| 15 | RTX 3060 laptopVast.ai | 0.04 USD/hr | 6 GB | 100 % | — | — | — | |
| 15 | Tesla V100Vast.ai | 0.04 USD/hr | 16 GB | 99.7 % | — | — | — | |
| 16 | Quadro P4000Vast.ai | 0.04 USD/hr | 8 GB | 99.8 % | — | — | — | |
| 17 | GTX 1080Vast.ai | 0.04 USD/hr | 8 GB | 99.9 % | — | — | — | |
| 17 | RTX 3060 TiVast.ai | 0.04 USD/hr | 8 GB | 99.7 % | — | — | — | |
| 18 | RTX 4090Vast.ai | 0.13 USD/hr | 22 GB | 99.9 % | — | — | — | |
| 19 | RTX 3060Vast.ai | 0.06 USD/hr | 12 GB | 100 % | — | — | — | |
| 20 | RTX PRO 5000Vast.ai | 0.6 USD/hr | 72 GB | 99.9 % | — | — | — | |
| 21 | A800 PCIEVast.ai | 0.47 USD/hr | 80 GB | 99.8 % | — | — | — | |
| 21 | H100 NVLVast.ai | 2 USD/hr | 94 GB | 99.2 % | — | — | — | |
| 22 | RTX 5060 TiVast.ai | 0.06 USD/hr | 8 GB | 99.8 % | — | — | — | |
| 23 | RTX 2060SVast.ai | 0.04 USD/hr | 8 GB | 99.6 % | — | — | — | |
| 24 | RTX 3090Vast.ai | 0.13 USD/hr | 24 GB | 100 % | — | — | — | |
| 25 | RTX A4000Vast.ai | 0.06 USD/hr | 16 GB | 99.9 % | — | — | — | |
| 26 | A100 SXM4Vast.ai | 0.27 USD/hr | 39 GB | 99.9 % | — | — | — | |
| 27 | RTX 5070Vast.ai | 0.06 USD/hr | 12 GB | 99.9 % | — | — | — | |
| 28 | RTX 2070Vast.ai | 0.05 USD/hr | 8 GB | 99.7 % | — | — | — | |
| 29 | GTX 1080 TiVast.ai | 0.06 USD/hr | 11 GB | 99.8 % | — | — | — | |
| 29 | H100 PCIEVast.ai | 1.8 USD/hr | 80 GB | 99.3 % | — | — | — | |
| 30 | GTX 1070 TiVast.ai | 0.05 USD/hr | 8 GB | 99.7 % | — | — | — | |
| 31 | RTX A6000Vast.ai | 0.29 USD/hr | 48 GB | 99.8 % | — | — | — | |
| 32 | RTX 5060Vast.ai | 0.06 USD/hr | 8 GB | 99.7 % | — | — | — | |
| 33 | Tesla P40Vast.ai | 0.07 USD/hr | 23 GB | 99.4 % | — | — | — | |
| 34 | RTX 5090Vast.ai | 0.21 USD/hr | 32 GB | 99.9 % | — | — | — | |
| 35 | RTX 4060Vast.ai | 0.06 USD/hr | 8 GB | 99.8 % | — | — | — | |
| 36 | A100 PCIEVast.ai | 0.27 USD/hr | 40 GB | 99.9 % | — | — | — | |
| 36 | Q RTX 8000Vast.ai | 0.27 USD/hr | 48 GB | 99.9 % | — | — | — | |
| 37 | RTX 4060 TiVast.ai | 0.08 USD/hr | 8 GB | 99.9 % | — | — | — | |
| 38 | RTX 3070 TiVast.ai | 0.07 USD/hr | 8 GB | 99.8 % | — | — | — | |
| 38 | RTX 4070S TiVast.ai | 0.07 USD/hr | 16 GB | 99.4 % | — | — | — | |
| 39 | Titan XpVast.ai | 0.06 USD/hr | 12 GB | 99.6 % | — | — | — | |
| 40 | RTX 6000AdaVast.ai | 0.54 USD/hr | 48 GB | 99.9 % | — | — | — | |
| 41 | L40SVast.ai | 0.47 USD/hr | 45 GB | 99.9 % | — | — | — | |
| 42 | Titan RTXVast.ai | 0.11 USD/hr | 24 GB | 99.9 % | — | — | — | |
| 43 | GTX 1660 TiVast.ai | 0.07 USD/hr | 6 GB | 99.8 % | — | — | — | |
| 44 | RTX 5880AdaVast.ai | 0.42 USD/hr | 48 GB | 99.8 % | — | — | — | |
| 44 | RTX 2070SVast.ai | 0.07 USD/hr | 8 GB | 99.7 % | — | — | — | |
| 45 | Tesla P100Vast.ai | 0.08 USD/hr | 16 GB | 99.5 % | — | — | — | |
| 46 | Q RTX 6000Vast.ai | 0.13 USD/hr | 23 GB | 99.9 % | — | — | — | |
| 47 | GTX 1660Vast.ai | 0.07 USD/hr | 6 GB | 99.7 % | — | — | — | |
| 48 | GTX 1070Vast.ai | 0.06 USD/hr | 8 GB | 99.3 % | — | — | — | |
| 49 | RTX 4070SVast.ai | 0.09 USD/hr | 12 GB | 99.6 % | — | — | — | |
| 50 | CMP 170HXVast.ai | 0.16 USD/hr | 40 GB | 99 % | — | — | — | |
| 50 | RTX 4080SVast.ai | 0.12 USD/hr | 16 GB | 99.7 % | — | — | — | |
| 51 | Titan VVast.ai | 0.09 USD/hr | 12 GB | 99.7 % | — | — | — | |
| 51 | Tesla T4Vast.ai | 0.12 USD/hr | 15 GB | 99.9 % | — | — | — | |
| 52 | RTX 4070 TiVast.ai | 0.09 USD/hr | 12 GB | 99.5 % | — | — | — | |
| 53 | RTX 4070Vast.ai | 0.08 USD/hr | 12 GB | 99.5 % | — | — | — | |
| 54 | RTX A5000Vast.ai | 0.23 USD/hr | 22 GB | 100 % | — | — | — | |
| 55 | RTX PRO 4000Vast.ai | 0.24 USD/hr | 24 GB | 99.7 % | — | — | — | |
| 56 | RTX 5000AdaVast.ai | 0.44 USD/hr | 32 GB | 99.9 % | — | — | — | |
| 57 | RTX 5070 TiVast.ai | 0.14 USD/hr | 16 GB | 99.5 % | — | — | — | |
| 58 | RTX 3050Vast.ai | 0.1 USD/hr | 8 GB | 99.7 % | — | — | — | |
| 59 | RTX PRO 4500Vast.ai | 0.31 USD/hr | 32 GB | 99.4 % | — | — | — | |
| 60 | A10Vast.ai | 0.24 USD/hr | 22 GB | 99.9 % | — | — | — | |
| 60 | RTX 3090 TiVast.ai | 0.14 USD/hr | 24 GB | 99 % | — | — | — | |
| 60 | RTX 5080Vast.ai | 0.17 USD/hr | 16 GB | 99.6 % | — | — | — | |
| 61 | RTX 4080Vast.ai | 0.12 USD/hr | 16 GB | 99.1 % | — | — | — | |
| 62 | RTX 3080 TiVast.ai | 0.1 USD/hr | 12 GB | 99 % | — | — | — | |
| 63 | RTX A4500Vast.ai | 0.11 USD/hr | 19 GB | 98.3 % | — | — | — | |
| 64 | L40Vast.ai | 0.4 USD/hr | 45 GB | 98.1 % | — | — | — | |
| 65 | RTX 4090DVast.ai | 0.44 USD/hr | 24 GB | 99 % | — | — | — | |
| 66 | Q RTX 5000Vast.ai | 0.08 USD/hr | 16 GB | 97.2 % | — | — | — | |
| 67 | Quadro GV100Vast.ai | 0.47 USD/hr | 32 GB | 97.1 % | — | — | — | |
| 68 | RTX 5060 laptopVast.ai | 0.14 USD/hr | 8 GB | 97.1 % | — | — | — |
classes
Datacenter GPUs
Enterprise accelerators with ECC memory and high interconnect bandwidth, sold for sustained training and large-batch inference.
18 GPU instances
Consumer GPUs
Gaming-class cards rented at lower hourly rates, typically on marketplace capacity, for fine-tuning and smaller inference workloads.
64 GPU instances
Serverless Inference
Per-second billed endpoints that scale to zero, where cold-start latency matters more than raw hourly rate.
0 GPU instances