Cloud GPU Prices per Hour

Live prices from 10 GPU clouds — How to read this →
H100 H200 A100 RTX 4090 Updated every few minutes

78 GPU models 2,628 listings 10 clouds Prices as of H100 from $1.98/h, A100 from $0.399/h, RTX 4090 from $0.340/h Data: GridPup

Cloud GPU rental prices in US dollars per GPU-hour
GPU Memory From /GPU/h Cheapest at Median 24h Clouds Details
Nvidia H100 Hopper · Data center 94 GB $1.98 GPU.ai $3.34 -7.7% 8
Nvidia A100 Ampere · Data center 80 GB $0.399 Vast.ai $1.31 +1.9% 8
Nvidia H200 Hopper · Data center 141 GB $3.59 Runpod $4.58 +0.0% 7
Nvidia B200 Blackwell · Data center 180 GB $6.10 Lium $8.39 +5.0% 6
RTX PRO 6000 Blackwell · Workstation 96 GB $1.33 Vast.ai $1.86 -10.8% 6
Nvidia L40S Ada Lovelace · Data center 48 GB $0.380 Lium $0.870 +0.0% 6
RTX A6000 Ampere · Workstation 48 GB $0.330 Runpod $0.450 -0.7% 5
Nvidia L40 Ada Lovelace · Data center 48 GB $0.330 Lium $0.780 +1.6% 5
RTX 4090 Ada Lovelace · Consumer 48 GB $0.340 Runpod $0.595 -20.1% 4
RTX 6000 Ada Ada Lovelace · Workstation 48 GB $0.453 Vast.ai $0.758 +0.0% 4
RTX 5090 Blackwell · Consumer 32 GB $0.480 Vast.ai $0.810 -17.4% 3
RTX 3090 Ampere · Consumer 24 GB $0.147 Vast.ai $0.243 -24.5% 3
RTX 5080 Blackwell · Consumer 16 GB $0.200 Vast.ai $0.360 +11.9% 3
Nvidia V100 Volta · Data center 32 GB $0.067 Vast.ai $0.210 +0.0% 3
RTX PRO 5000 Blackwell · Workstation 72 GB $0.560 Vast.ai $0.820 +0.0% 3
RTX PRO 4000 Blackwell · Workstation 24 GB $0.227 Vast.ai $0.448 +7.0% 2
Nvidia L4 Ada Lovelace · Data center 24 GB $0.333 Vast.ai $0.493 -15.7% 3
Nvidia B300 Blackwell Ultra · Data center 288 GB $8.99 Runpod $10.00 +11.2% 3
RTX 5060 Ti Blackwell · Consumer 16 GB $0.107 Vast.ai $0.133 -27.3% 1
RTX 5070 Ti Blackwell · Consumer 16 GB $0.157 Vast.ai $0.245 +2.8% 2
RTX 3080 Ampere · Consumer 20 GB $0.107 Vast.ai $0.157 -4.5% 2
RTX 3080 Ti Ampere · Consumer 12 GB $0.107 Vast.ai $0.173 +2.0% 2
RTX 4070 Ti SUPER Ada Lovelace · Consumer 16 GB $0.133 Vast.ai $0.240 +2.8% 1
RTX A4000 Ampere · Workstation 16 GB $0.079 Vast.ai $0.151 +25.8% 2
RTX 3070 Ampere · Consumer 8 GB $0.067 Vast.ai $0.112 +0.0% 2
RTX 4080 Ada Lovelace · Consumer 32 GB $0.240 Vast.ai $0.265 +0.0% 2
RTX 4000 Ada Ada Lovelace · Workstation 20 GB $0.173 Vast.ai $0.210 +1.6% 2
Nvidia A40 Ampere · Data center 48 GB $0.267 Vast.ai $0.502 -3.8% 2
RTX 4080 SUPER Ada Lovelace · Consumer 16 GB $0.233 Vast.ai $0.277 -5.7% 2
RTX 3090 Ti Ampere · Consumer 24 GB $0.243 Vast.ai $0.268 +0.6% 2
RTX 4070 Ti Ada Lovelace · Consumer 12 GB $0.133 Vast.ai $0.168 -12.9% 2
RTX A2000 Ampere · Workstation 12 GB $0.051 Vast.ai $0.133 -1.2% 2
RTX 5000 Ada Ada Lovelace · Workstation 32 GB $0.490 Runpod $1.31 +0.0% 2
RTX A5000 Ampere · Workstation 24 GB $0.160 Runpod $0.225 +1.1% 2
RTX PRO 4500 Blackwell · Workstation 32 GB $0.347 Vast.ai $0.533 -25.9% 2
RTX A4500 Ampere · Workstation 20 GB $0.190 Runpod $0.220 -14.1% 1
AMD MI300X CDNA 3 · Data center 192 GB $2.39 Runpod $2.39 +0.0% 1
RTX 3060 Ampere · Consumer 12 GB $0.051 Vast.ai $0.079 +0.0% 1
RTX 4060 Ti Ada Lovelace · Consumer 16 GB $0.107 Vast.ai $0.137 -14.6% 1
Titan Xp Pascal · Consumer 12 GB $0.060 Vast.ai $0.067 +0.0% 1
RTX 5070 Blackwell · Consumer 12 GB $0.133 Vast.ai $0.240 +12.5% 1
RTX 2080 Ti Turing · Consumer 22 GB $0.067 Vast.ai $0.127 +6.7% 1
RTX 4070 Ada Lovelace · Consumer 12 GB $0.120 Vast.ai $0.160 -11.1% 1
RTX 3060 Ti Ampere · Consumer 8 GB $0.065 Vast.ai $0.093 +0.0% 1
RTX 4070 SUPER Ada Lovelace · Consumer 12 GB $0.107 Vast.ai $0.187 +3.7% 1
GTX 1070 Ti Pascal · Consumer 8 GB $0.051 Vast.ai $0.053 -7.0% 1
RTX 2060 Turing · Consumer 12 GB $0.053 Vast.ai $0.093 +55.5% 1
GTX 1080 Pascal · Consumer 8 GB $0.051 Vast.ai $0.053 +0.0% 1
RTX 2070 SUPER Turing · Consumer 8 GB $0.060 Vast.ai $0.077 +0.0% 1
RTX 4060 Ada Lovelace · Consumer 8 GB $0.073 Vast.ai $0.079 -0.8% 1
Titan RTX Turing · Consumer 24 GB $0.200 Vast.ai $0.236 -1.7% 1
GTX 1080 Ti Pascal · Consumer 11 GB $0.060 Vast.ai $0.063 -3.1% 1
GTX 1660 SUPER Turing · Consumer 6 GB $0.056 Vast.ai $0.060 +3.5% 1
RTX 5880 Ada Ada Lovelace · Workstation 48 GB $0.800 Vast.ai $0.800 +0.0% 1
Nvidia A30 Ampere · Data center 24 GB $0.410 JarvisLabs $0.410 +0.0% 1
Quadro P4000 Pascal · Workstation 8 GB $0.053 Vast.ai $0.070 +0.0% 1
RTX 5060 Blackwell · Consumer 8 GB $0.093 Vast.ai $0.117 +0.0% 1
GTX 1660 Ti Turing · Consumer 6 GB $0.060 Vast.ai $0.067 +0.0% 1
Nvidia P100 Pascal · Data center 16 GB $0.080 Vast.ai $0.080 -14.3% 1
Quadro RTX 8000 Turing · Workstation 48 GB $0.253 Vast.ai $0.253 -5.0% 1
RTX 2070 Turing · Consumer 8 GB $0.053 Vast.ai $0.060 +0.0% 1
RTX 3070 Ti Ampere · Consumer 8 GB $0.160 Vast.ai $0.160 +0.0% 1
Nvidia A10 Ampere · Data center 24 GB $0.240 Vast.ai $0.240 +0.0% 1
Nvidia P40 Pascal · Data center 24 GB $0.105 Vast.ai $0.105 +0.0% 1
RTX 2060 SUPER Turing · Consumer 8 GB $0.069 Vast.ai $0.085 +12.3% 1
RTX 3050 Ampere · Consumer — $0.067 Vast.ai $0.121 — 1
RTX PRO 6000D Blackwell · Workstation 84 GB $1.47 Vast.ai $1.47 +0.0% 1
Titan V Volta · Consumer 12 GB $0.160 Vast.ai $0.167 -16.7% 1
Quadro RTX 6000 Turing · Workstation 24 GB $0.147 Vast.ai $0.210 +37.0% 1
GTX 1070 Pascal · Consumer 8 GB $0.400 Vast.ai $0.400 +756.5% 1
GTX 1660 Turing · Consumer 6 GB $0.667 Vast.ai $0.667 — 1
GTX Titan X Maxwell · Consumer 12 GB $0.067 Vast.ai $0.067 +0.0% 1
Quadro RTX 5000 Turing · Workstation 16 GB $0.093 Vast.ai $0.093 +0.0% 1
RTX 4090D Ada Lovelace · Consumer 48 GB $0.400 Vast.ai $0.400 — 1
RTX 4500 Ada Ada Lovelace · Workstation 24 GB $0.387 Vast.ai $0.387 -9.4% 1
RTX 2000 Ada Ada Lovelace · Workstation 16 GB $0.240 Runpod $0.240 +0.0% 1
RTX PRO 6000 MIG 1g.24gb Blackwell · Workstation 24 GB $0.690 Runpod $0.690 +0.0% 1
RTX PRO 6000 MIG 2g.48gb Blackwell · Workstation 48 GB $1.29 Runpod $1.29 +0.0% 1

On-demand prices in US dollars per GPU-hour, before tax. Machines with several GPUs are divided by their GPU count. Prices and availability by GridPup, the cloud GPU price comparison.

How to read cloud GPU prices

Each row is one GPU model. From is the lowest on-demand price listed right now; Median is the middle price across every listing, which is closer to what you will usually pay. 24h shows how the cheapest price moved over the last day.

Choosing a GPU

Why prices differ between clouds

Marketplace clouds rent spare machines from many hosts, so prices are low but vary by host and reliability. Managed clouds run their own data centers and charge more for fixed capacity and support. Storage and data transfer are billed separately on most clouds.

Training, benchmarking and deploying

Once you know the GPU, you can train or fine-tune a model on it from your browser, benchmark an LLM's tokens per second on real GPUs, check how much GPU memory an open model needs, deploy it as an API and test its latency, or compare LLMs side by side in the AI Playground.

Where the data comes from

Prices are collected every few minutes by GridPup, which compares GPU clouds and lets you rent from them in one place. Open a GPU to see each listing, its location and price history, or compare GPUs side by side.

Support This Free Tool

Every coffee helps keep the servers running. Every book sale funds the next tool I'm dreaming up. You're not just supporting a site — you're helping me build what developers actually need.

500K+ users
200+ tools
100% private
☕ One-time support Buy me a coffee 📚 Learn & support 9-Book Bundle - $9 𝕏 Stay updated Follow @anish2good
Privacy Guarantee: Private keys you enter or generate are never stored on our servers. All tools are served over HTTPS.