GPUs with 16 GB+ VRAM
32 GPU models with at least 16 GB, ranked by on-demand price when available and otherwise by the cheapest listed tier. VRAM is the binding constraint for most AI workloads: if the model does not fit, the price does not matter.
| GPU | VRAM | On-demand | Spot/community from | $/hr per GB | Where | |
|---|---|---|---|---|---|---|
| RTX A4000 | 16 GB | $0.10 | $0.17 | $0.01 | TensorDock on-demand | Rent → |
| Tesla V100 | 16 GB | $0.17 | $0.19 | $0.01 | DataCrunch on-demand | Rent → |
| RTX 5080 | 16 GB | community cloud only | $0.18 | $0.01 | SaladCloud community cloud | Rent → |
| RTX 3090 | 24 GB | $0.20 | $0.09 | $0.01 | TensorDock on-demand | Rent → |
| RTX A5000 | 24 GB | $0.27 | $0.16 | $0.01 | RunPod secure cloud | Rent → |
| Tesla T4 | 16 GB | $0.35 | $0.19 | $0.02 | Google Cloud on-demand | Rent → |
| A30 | 24 GB | $0.35 | – | $0.01 | Massed Compute on-demand | Rent → |
| RTX A6000 | 48 GB | $0.35 | $0.33 | $0.01 | Thunder Compute on-demand | Rent → |
| RTX 4090 | 24 GB | $0.35 | $0.16 | $0.01 | TensorDock on-demand | Rent → |
| L4 | 24 GB | $0.44 | $0.10 | $0.02 | Jarvislabs on-demand | Rent → |
| A40 | 48 GB | $0.44 | – | $0.01 | RunPod secure cloud | Rent → |
| Quadro RTX 5000 | 16 GB | $0.60 | – | $0.04 | OVHcloud on-demand | Rent → |
| Quadro RTX 6000 | 24 GB | $0.69 | – | $0.03 | Lambda on-demand | Rent → |
| RTX PRO 4500 | 32 GB | $0.72 | – | $0.02 | RunPod secure cloud | Rent → |
| RTX 6000 Ada | 48 GB | $0.75 | $0.74 | $0.02 | TensorDock on-demand | Rent → |
| L40 | 48 GB | $0.79 | $0.69 | $0.02 | Thunder Compute on-demand | Rent → |
| RTX 5090 | 32 GB | $0.86 | $0.25 | $0.03 | Spheron on-demand | Rent → |
| L40S | 48 GB | $0.88 | $0.79 | $0.02 | Massed Compute on-demand | Rent → |
| A100 | 80 GB | $0.89 | $0.68 | $0.01 | Jarvislabs on-demand | Rent → |
| A10 | 24 GB | $1.00 | – | $0.04 | OVHcloud on-demand | Rent → |
| RTX PRO 6000 | 96 GB | $1.10 | $0.48 | $0.01 | Google Cloud on-demand | Rent → |
| H100 | 94 GB | $1.99 | $0.94 | $0.02 | Voltage Park on-demand | Rent → |
| GH200 | 96 GB | $2.29 | – | $0.02 | Lambda on-demand | Rent → |
| MI300X | 192 GB | $2.39 | – | $0.01 | RunPod secure cloud | Rent → |
| MI355X | 288 GB | spot / interruptible only | $2.59 | $0.01 | Vultr spot / interruptible | Rent → |
| H200 | 143 GB | $2.60 | $2.51 | $0.02 | GMI Cloud on-demand | Rent → |
| B200 | 180 GB | $3.49 | $3.20 | $0.02 | Prime Intellect on-demand | Rent → |
| MI325X | 256 GB | $3.80 | $2.00 | $0.01 | DigitalOcean on-demand | Rent → |
| B300 | 288 GB | $4.99 | $4.70 | $0.02 | Prime Intellect on-demand | Rent → |
| MI350X | 288 GB | $6.16 | – | $0.02 | DigitalOcean on-demand | Rent → |
| GB200 | 186 GB | $8.00 | – | $0.04 | GMI Cloud on-demand | Rent → |
| GB300 | 288 GB | $8.62 | – | $0.03 | DataCrunch on-demand | Rent → |
Tables rank by price, lowest first. Commercial relationships never affect ordering.How we collect and rank prices.
FAQ
What is the cheapest 16 GB+ GPU to rent?
The cheapest is the RTX A4000 (16 GB) at $0.10 per GPU-hour on TensorDock (on-demand), verified 2026-08-22.
How many GPUs have at least 16 GB of VRAM?
32 GPU models in our index have at least 16 GB of VRAM.
Other sizes: 16 GB+ · 24 GB+ · 32 GB+ · 48 GB+ · 80 GB+ · 96 GB+ · 140 GB+ · 192 GB+ · all GPUs · full table.