Cheapest cloud GPU prices right now
The lowest on-demand price per GPU-hour for every GPU model we track across 30 cloud providers, next to the cheapest option that reported stock on our last hourly check. Ordered by the in-stock price, cheapest first. The cheapest H100 in stock right now is $0.41/GPU-hr on Vast (1× node).
Stock last verified · 61 of 83 models have at least one in-stock listing.
Cheapest in-stock price for the popular GPUs
Every GPU model, cheapest in-stock price first
“Cheapest” is the lowest listed on-demand price per GPU-hour; “cheapest in stock” is the lowest from a provider that reported capacity on a check in the last 6 hours. Models with nothing in stock sit at the bottom. Click a GPU for every provider’s price, stock and 30-day reliability.
| GPU | Cheapest / GPU-hr | Cheapest in stock | Stock · 30d | Providers | Spot / GPU-hr |
|---|---|---|---|---|---|
| RTX 206012 GB | $0.01Vast · 3× | $0.01Vast · 3× | Limited0% · 30d | 1 | — |
| L422 GB | $0.02Vast · 1× | $0.02Vast · 1× | Limited47% · 30d | 6 | — |
| V10016 GB | $0.02Vast · 1× | $0.02Vast · 1× | Limited82% · 30d | 14 | $0.02 |
| P48 GB | $0.02Vast · 1× | $0.02Vast · 1× | Limited0% · 30d | 2 | $0.01 |
| RTX A20006 GB | $0.03Vast · 2× | $0.03Vast · 2× | Limited9% · 30d | 2 | $0.03 |
| RTX 306012 GB | $0.02Vast · 1× | $0.04Vast · 1× | Limited94% · 30d | 2 | $0.01 |
| RTX 3060 Ti8 GB | $0.03Vast · 1× | $0.04Vast · 1× | Limited94% · 30d | 1 | $0.02 |
| RTX 2080 Ti11 GB | $0.04Vast · 2× | $0.04Vast · 1× | Limited94% · 30d | 1 | $0.03 |
| RTX 30708 GB | $0.04Vast · 1× | $0.04Vast · 1× | Limited94% · 30d | 4 | — |
| RTX 20708 GB | $0.05Vast · 4× | $0.05Vast · 4× | Limited0% · 30d | 1 | $0.03 |
| RTX 40608 GB | $0.05Vast · 8× | $0.05Vast · 8× | Limited0% · 30d | 2 | — |
| RTX 4060 Ti16 GB | $0.05Vast · 6× | $0.05Vast · 6× | Limited0% · 30d | 2 | $0.04 |
| RTX 5060 Ti16 GB | $0.04Vast · 2× | $0.05Vast · 1× | Limited94% · 30d | 1 | $0.03 |
| RTX 3070 Ti8 GB | $0.05Vast · 1× | $0.06Vast · 1× | Limited91% · 30d | 1 | $0.04 |
| RTX 50608 GB | $0.06Vast · 2× | $0.06Vast · 2× | Limited0% · 30d | 1 | — |
| RTX A400016 GB | $0.05Vast · 5× | $0.06Vast · 2× | Limited94% · 30d | 5 | $0.04 |
| RTX 308010 GB | $0.05Vast · 1× | $0.07Vast · 6× | Limited0% · 30d | 3 | $0.03 |
| RTX 4070 Ti16 GB | $0.07Vast · 2× | $0.07Vast · 2× | Limited88% · 30d | 2 | $0.07 |
| RTX 407012 GB | $0.07Vast · 4× | $0.07Vast · 2× | Limited90% · 30d | 2 | $0.07 |
| Titan V12 GB | $0.09Vast · 2× | $0.09Vast · 2× | Limited0% · 30d | 1 | $0.07 |
| RTX 3080 Ti12 GB | $0.05Vast · 1× | $0.09Vast · 3× | Limited0% · 30d | 3 | $0.03 |
| RTX 309024 GB | $0.05Vast · 8× | $0.09Vast · 8× | Limited15% · 30d | 4 | — |
| Tesla P4023 GB | $0.03Vast · 2× | $0.11Vast · 2× | Limited0% · 30d | 1 | $0.03 |
| Titan RTX24 GB | $0.11Vast · 4× | $0.11Vast · 4× | Limited0% · 30d | 1 | $0.09 |
| RTX 408016 GB | $0.10Vast · 1× | $0.11Vast · 1× | Limited94% · 30d | 3 | $0.09 |
| RTX A450020 GB | $0.08Vast · 1× | $0.11Vast · 1× | Limited0% · 30d | 2 | $0.07 |
| RTX 5070 Ti16 GB | $0.08Vast · 1× | $0.11Vast · 2× | Limited81% · 30d | 1 | $0.07 |
| RTX 3090 Ti24 GB | $0.12Vast · 1× | $0.13Vast · 1× | Limited | 2 | $0.09 |
| Quadro RTX 600023 GB | $0.13Vast · 1× | $0.13Vast · 8× | Limited0% · 30d | 1 | $0.11 |
| RTX 409024 GB | $0.13Vast · 8× | $0.13Vast · 8× | Limited88% · 30d | 6 | — |
| RTX 508016 GB | $0.10Vast · 4× | $0.13Vast · 2× | Limited94% · 30d | 3 | $0.09 |
| RTX A500024 GB | $0.07Vast · 1× | $0.16Runpod · 1× | Limited0% · 30d | 3 | $0.07 |
| RTX PRO 400024 GB | $0.16Vast · 2× | $0.16Vast · 2× | Limited | 2 | — |
| RTX 4070 SUPER12 GB | $0.17Theta EdgeCloud · 1× | $0.17Theta EdgeCloud · 1× | Limited0% · 30d | 1 | — |
| RTX 4000 Ada20 GB | $0.14Vast · 1× | $0.19Vast · 1× | Limited0% · 30d | 3 | $0.13 |
| A1022 GB | $0.20Vast · 4× | $0.20Vast · 1× | Limited0% · 30d | 5 | — |
| RTX 2000 Ada16 GB | $0.24Runpod · 1× | $0.24Runpod · 1× | Limited0% · 30d | 1 | — |
| RTX 4070 Ti SUPER16 GB | $0.25Theta EdgeCloud · 1× | $0.25Theta EdgeCloud · 1× | Limited0% · 30d | 1 | — |
| RTX 507012 GB | $0.25Theta EdgeCloud · 1× | $0.25Theta EdgeCloud · 1× | In stock57% · 30d | 1 | — |
| A10040 GB | $0.20Vast · 1× | $0.27Vast · 2× | Limited53% · 30d | 13 | $0.10 |
| Quadro RTX 800045 GB | $0.25Vast · 4× | $0.27Vast · 1× | Limited0% · 30d | 1 | — |
| RTX 509032 GB | $0.27Vast · 8× | $0.29Vast · 4× | Limited94% · 30d | 6 | $0.11 |
| RTX PRO 450032 GB | $0.29Vast · 1× | $0.29Vast · 1× | Limited2% · 30d | 3 | $0.10 |
| RTX A600048 GB | $0.29Vast · 1× | $0.31Vast · 2× | Limited6% · 30d | 8 | — |
| L40180 GB | $0.33Lium · 4× | $0.33Lium · 4× | Limited | 6 | — |
| A402 GB | $0.07Vultr · 1× | $0.34Vast · 1× | Limited | 4 | — |
| L40S90 GB | $0.38Lium · 2× | $0.38Lium · 2× | In stock40% · 30d | 14 | — |
| RTX 6000 Ada48 GB | $0.28Vast · 1× | $0.39Vast · 1× | Limited35% · 30d | 6 | $0.27 |
| H10080 GB | $0.41Vast · 4× | $0.41Vast · 1× | Limited89% · 30d | 21 | $0.33 |
| RTX 4080 SUPER16 GB | $0.44Theta EdgeCloud · 1× | $0.44Theta EdgeCloud · 1× | Limited0% · 30d | 1 | — |
| A100-80GB80 GB | $0.44Vast · 1× | $0.47Vast · 2× | Limited94% · 30d | 16 | $0.13 |
| A162 GB | $0.06Vultr · 1× | $0.47Vultr · 2× | In stock100% · 30d | 3 | — |
| RTX 5000 Ada32 GB | $0.49Runpod · 1× | $0.49Runpod · 1× | Limited0% · 30d | 1 | — |
| RTX PRO 500048 GB | $0.40Vast · 1× | $0.57Vast · 1× | Limited68% · 30d | 2 | $0.31 |
| RTX 20808 GB | $0.74Vast · 1× | $0.74Vast · 1× | Limited0% · 30d | 1 | $0.03 |
| RTX PRO 600096 GB | $0.66Vast · 1× | $0.78Vast · 1× | Limited94% · 30d | 10 | — |
| RTX 600024 GB | $0.29Seeweb · 1× | $1.80Nebius · 1× | Limited3% · 30d | 3 | — |
| H200564 GB | $2.00PrimeIntellect · 4× | $2.00PrimeIntellect · 8× | In stock100% · 30d | 14 | — |
| MI300X192 GB | $2.39Runpod · 1× | $2.39Runpod · 1× | Limited0% · 30d | 2 | — |
| B2001440 GB | $4.40PrimeIntellect · 8× | $4.49Vast · 1× | Limited1% · 30d | 10 | — |
| B300288 GB | $6.94Runpod · 1× | $7.85Nebius · 1× | Limited0% · 30d | 5 | — |
| K80 | $0.18Azure · 1× | none in stock | No data | 1 | — |
| RTX 4000 SFF Ada20 GB | $0.18Runpod · 1× | none in stock | Out0% · 30d | 1 | — |
| T416 GB | $0.20Theta EdgeCloud · 1× | none in stock | No data100% · 30d | 4 | — |
| Inferentia8 GB | $0.23AWS · 1× | none in stock | No data | 1 | $0.04 |
| M60 | $0.23Azure · 1× | none in stock | No data | 1 | — |
| Radeon MI25 | $0.23Azure · 1× | none in stock | No data | 1 | $0.03 |
| A3024 GB | $0.29Seeweb · 1× | none in stock | No data | 2 | — |
| Radeon Pro V5208 GB | $0.38AWS · 1× | none in stock | No data | 1 | $0.12 |
| T4g16 GB | $0.42AWS · 1× | none in stock | No data | 1 | $0.14 |
| Inferentia232 GB | $0.76AWS · 1× | none in stock | No data | 1 | $0.08 |
| A10G22 GB | $1.01AWS · 1× | none in stock | No data | 1 | $0.48 |
| GAUDI2 | $1.10Shadeform · 8× | none in stock | No data | 1 | — |
| P10016 GB | $1.24Scaleway · 1× | none in stock | No data | 3 | — |
| RTX PRO 6000 96GB | $1.25Seeweb · 1× | none in stock | No data | 1 | — |
| Trainium32 GB | $1.34AWS · 1× | none in stock | No data | 1 | $0.20 |
| TPU v5 Lite | $1.96Google Cloud · 1× | none in stock | No data | 1 | $0.70 |
| Virtex UltraScale+ (VU47P) | $1.98AWS · 1× | none in stock | No data | 1 | $0.74 |
| GH20096 GB | $2.29Lambda · 1× | none in stock | Out41% · 30d | 2 | — |
| TPU v6e | $3.46Google Cloud · 1× | none in stock | No data | 1 | $1.01 |
| TPU v2 | $5.26Google Cloud · 1× | none in stock | No data | 1 | $1.71 |
| TPU v3 | $8.76Google Cloud · 1× | none in stock | No data | 1 | — |
| TPU v5p | $17.56Google Cloud · 1× | none in stock | No data | 1 | $3.61 |
Frequently asked questions
What is the cheapest cloud GPU right now?
The table on this page is ordered by the cheapest in-stock on-demand price per GPU-hour, so the top row is the cheapest GPU you can start on right now. For a given model, the "Cheapest" column is the lowest listed price and the "Cheapest in stock" column is the lowest price from a provider that reported capacity on a check in the last 6 hours — that second number is the one you can actually rent.
Why is the cheapest listing not always in stock?
Marketplace and spot-style providers list rates for machines that come and go. We poll stock hourly from provider APIs; a listing that is cheap but out of stock is shown for reference, and the in-stock column tells you the real floor you can start on now.
How is the price per GPU-hour calculated?
Node price divided by the number of GPUs in the node. An 8× H100 node at $23.92/hr is $2.99 per GPU-hour, which is how it is compared against a 1× H100 at $2.49/hr. Prices are on-demand USD per hour as published by each provider; billing increments and egress fees are listed on the provider pages.
How do I rent the cheapest cloud GPU?
Click the GPU model for every provider's price and stock, or the provider name for its full catalogue. Every row links to the provider through our stable /go links; a few carry a disclosed referral code that never affects ranking or price.