# gpu.compare > Cloud GPU pricing is scattered across provider catalogs that don't speak the same language. This site puts GPU pricing, geography, and provider-reported availability in one place. It tracks providers hourly and keeps the history, so it answers both what is cheapest and what changed. Updated hourly. Last scan 2026-11-16T22:30:26.764Z. The priced slice covers 4020 instances from 10 providers across 155 regions. Every page listed below has a Markdown twin at the same path plus `.md`, which is the cheapest way to read one. Those pages also return their twin at the HTML URL when the request sends `Accept: text/markdown`. Offering pages and filtered URLs do not, so read those at their `.md` URLs or through /catalog.md. Use the Markdown catalog to filter and paginate offerings. ## Pages - [Compare cloud GPU prices and availability](https://gpu.compare/index.md): Compare on-demand GPU prices across cloud providers. Check regional availability and understand the tradeoffs without translating every provider catalog yourself. Canonical HTML page: https://gpu.compare/ - [Compare cloud providers](https://gpu.compare/providers.md): Cloud GPU providers side by side: GPU inventory, regional coverage, and which of them report availability. Each provider also has its own page under /providers/. Canonical HTML page: https://gpu.compare/providers - [How we compare GPU prices](https://gpu.compare/methodology.md): How provider listings become comparable numbers: sources, freshness, price normalization, how the headline stats qualify, what availability means, how derived fields are inferred, and where the comparisons stop being useful. Canonical HTML page: https://gpu.compare/methodology ## GPU model pages One page per accelerator, each answering the price of renting that model by the hour. They do not cover the cost of buying the card. A model appears here only when more than one provider sells it, because a single-provider listing is not a comparison; the rest stay reachable as `q` searches on /catalog.md. - [NVIDIA H100 cloud prices and availability](https://gpu.compare/gpu/h100.md): Compare every NVIDIA H100 in the catalog by provider, region, instance size, purchase model, price per GPU-hour, and provider-reported availability. Prices are for renting by the hour, not for buying the card. Canonical HTML page: https://gpu.compare/gpu/h100 - [NVIDIA H200 cloud prices and availability](https://gpu.compare/gpu/h200.md): Compare every NVIDIA H200 in the catalog by provider, region, instance size, purchase model, price per GPU-hour, and provider-reported availability. Prices are for renting by the hour, not for buying the card. Canonical HTML page: https://gpu.compare/gpu/h200 - [NVIDIA A100 cloud prices and availability](https://gpu.compare/gpu/a100.md): Compare every NVIDIA A100 in the catalog by provider, region, instance size, purchase model, price per GPU-hour, and provider-reported availability. Prices are for renting by the hour, not for buying the card. Canonical HTML page: https://gpu.compare/gpu/a100 - [NVIDIA B200 cloud prices and availability](https://gpu.compare/gpu/b200.md): Compare every NVIDIA B200 in the catalog by provider, region, instance size, purchase model, price per GPU-hour, and provider-reported availability. Prices are for renting by the hour, not for buying the card. Canonical HTML page: https://gpu.compare/gpu/b200 - [NVIDIA B300 cloud prices and availability](https://gpu.compare/gpu/b300.md): Compare every NVIDIA B300 in the catalog by provider, region, instance size, purchase model, price per GPU-hour, and provider-reported availability. Prices are for renting by the hour, not for buying the card. Canonical HTML page: https://gpu.compare/gpu/b300 - [NVIDIA RTX PRO 6000 cloud prices and availability](https://gpu.compare/gpu/rtx-pro-6000.md): Compare every NVIDIA RTX PRO 6000 in the catalog by provider, region, instance size, purchase model, price per GPU-hour, and provider-reported availability. Prices are for renting by the hour, not for buying the card. Canonical HTML page: https://gpu.compare/gpu/rtx-pro-6000 - [NVIDIA RTX 6000 Ada cloud prices and availability](https://gpu.compare/gpu/rtx-6000-ada.md): Compare every NVIDIA RTX 6000 Ada in the catalog by provider, region, instance size, purchase model, price per GPU-hour, and provider-reported availability. Prices are for renting by the hour, not for buying the card. Canonical HTML page: https://gpu.compare/gpu/rtx-6000-ada - [NVIDIA RTX A6000 cloud prices and availability](https://gpu.compare/gpu/rtx-a6000.md): Compare every NVIDIA RTX A6000 in the catalog by provider, region, instance size, purchase model, price per GPU-hour, and provider-reported availability. Prices are for renting by the hour, not for buying the card. Canonical HTML page: https://gpu.compare/gpu/rtx-a6000 - [NVIDIA L40S cloud prices and availability](https://gpu.compare/gpu/l40s.md): Compare every NVIDIA L40S in the catalog by provider, region, instance size, purchase model, price per GPU-hour, and provider-reported availability. Prices are for renting by the hour, not for buying the card. Canonical HTML page: https://gpu.compare/gpu/l40s - [NVIDIA RTX 4000 Ada cloud prices and availability](https://gpu.compare/gpu/rtx-4000-ada.md): Compare every NVIDIA RTX 4000 Ada in the catalog by provider, region, instance size, purchase model, price per GPU-hour, and provider-reported availability. Prices are for renting by the hour, not for buying the card. Canonical HTML page: https://gpu.compare/gpu/rtx-4000-ada ## GPU comparisons Head-to-head pages for accelerator pairs with measured search demand, plus pairs a model page recommends. Each states which of the two is cheaper per GPU-hour and per GB of VRAM, and how often either was purchasable. None of them benchmarks the two: this site prices compute and does not measure throughput, and an hourly price is no substitute for that. - [NVIDIA H100 vs NVIDIA A100](https://gpu.compare/gpu/compare/h100-vs-a100.md): What renting an H100 costs against an A100, per GPU-hour and per GB of VRAM, and how often each was in stock. Prices only: this page doesn't benchmark the two. Canonical HTML page: https://gpu.compare/gpu/compare/h100-vs-a100 - [NVIDIA H100 vs NVIDIA H200](https://gpu.compare/gpu/compare/h100-vs-h200.md): What renting an H100 costs against an H200, per GPU-hour and per GB of VRAM, and how often each was in stock. Prices only: this page doesn't benchmark the two. Canonical HTML page: https://gpu.compare/gpu/compare/h100-vs-h200 - [NVIDIA A100 vs NVIDIA L40S](https://gpu.compare/gpu/compare/a100-vs-l40s.md): What renting an A100 costs against an L40S, per GPU-hour and per GB of VRAM, and how often each was in stock. Prices only: this page doesn't benchmark the two. Canonical HTML page: https://gpu.compare/gpu/compare/a100-vs-l40s - [NVIDIA A100 vs NVIDIA RTX A6000](https://gpu.compare/gpu/compare/a100-vs-rtx-a6000.md): What renting an A100 costs against an RTX A6000, per GPU-hour and per GB of VRAM, and how often each was in stock. Prices only: this page doesn't benchmark the two. Canonical HTML page: https://gpu.compare/gpu/compare/a100-vs-rtx-a6000 - [NVIDIA B200 vs NVIDIA H100](https://gpu.compare/gpu/compare/b200-vs-h100.md): What renting a B200 costs against an H100, per GPU-hour and per GB of VRAM, and how often each was in stock. Prices only: this page doesn't benchmark the two. Canonical HTML page: https://gpu.compare/gpu/compare/b200-vs-h100 - [NVIDIA H100 vs NVIDIA RTX PRO 6000](https://gpu.compare/gpu/compare/h100-vs-rtx-pro-6000.md): What renting an H100 costs against an RTX PRO 6000, per GPU-hour and per GB of VRAM, and how often each was in stock. Prices only: this page doesn't benchmark the two. Canonical HTML page: https://gpu.compare/gpu/compare/h100-vs-rtx-pro-6000 - [NVIDIA H200 vs NVIDIA B200](https://gpu.compare/gpu/compare/h200-vs-b200.md): What renting an H200 costs against a B200, per GPU-hour and per GB of VRAM, and how often each was in stock. Prices only: this page doesn't benchmark the two. Canonical HTML page: https://gpu.compare/gpu/compare/h200-vs-b200 - [NVIDIA A100 vs NVIDIA RTX PRO 6000](https://gpu.compare/gpu/compare/a100-vs-rtx-pro-6000.md): What renting an A100 costs against an RTX PRO 6000, per GPU-hour and per GB of VRAM, and how often each was in stock. Prices only: this page doesn't benchmark the two. Canonical HTML page: https://gpu.compare/gpu/compare/a100-vs-rtx-pro-6000 - [NVIDIA B200 vs NVIDIA B300](https://gpu.compare/gpu/compare/b200-vs-b300.md): What renting a B200 costs against a B300, per GPU-hour and per GB of VRAM, and how often each was in stock. Prices only: this page doesn't benchmark the two. Canonical HTML page: https://gpu.compare/gpu/compare/b200-vs-b300 - [NVIDIA RTX 6000 Ada vs NVIDIA RTX A6000](https://gpu.compare/gpu/compare/rtx-6000-ada-vs-rtx-a6000.md): What renting an RTX 6000 Ada costs against an RTX A6000, per GPU-hour and per GB of VRAM, and how often each was in stock. Prices only: this page doesn't benchmark the two. Canonical HTML page: https://gpu.compare/gpu/compare/rtx-6000-ada-vs-rtx-a6000 - [NVIDIA RTX 6000 Ada vs NVIDIA L40S](https://gpu.compare/gpu/compare/rtx-6000-ada-vs-l40s.md): What renting an RTX 6000 Ada costs against an L40S, per GPU-hour and per GB of VRAM, and how often each was in stock. Prices only: this page doesn't benchmark the two. Canonical HTML page: https://gpu.compare/gpu/compare/rtx-6000-ada-vs-l40s ## Provider pages One page per provider, with GPU prices, inventory, regions, and availability reporting. Three classes exist: providers that publish both available and unavailable, providers that publish nothing, and providers that only ever say available. The last two cannot confirm that a price is purchasable. - [AWS GPU pricing and availability](https://gpu.compare/providers/aws.md): AWS GPU prices per GPU-hour, inventory by region, and how AWS reports availability. Canonical HTML page: https://gpu.compare/providers/aws - [Azure GPU pricing and availability](https://gpu.compare/providers/azure.md): Azure GPU prices per GPU-hour, inventory by region, and how Azure reports availability. Canonical HTML page: https://gpu.compare/providers/azure - [DigitalOcean GPU pricing and availability](https://gpu.compare/providers/digitalocean.md): DigitalOcean GPU prices per GPU-hour, inventory by region, and how DigitalOcean reports availability. Canonical HTML page: https://gpu.compare/providers/digitalocean - [Google Cloud GPU pricing and availability](https://gpu.compare/providers/gcp.md): Google Cloud GPU prices per GPU-hour, inventory by region, and how Google Cloud reports availability. Canonical HTML page: https://gpu.compare/providers/gcp - [Hyperstack GPU pricing and availability](https://gpu.compare/providers/hyperstack.md): Hyperstack GPU prices per GPU-hour, inventory by region, and how Hyperstack reports availability. Canonical HTML page: https://gpu.compare/providers/hyperstack - [Lambda GPU pricing and availability](https://gpu.compare/providers/lambda.md): Lambda GPU prices per GPU-hour, inventory by region, and how Lambda reports availability. Canonical HTML page: https://gpu.compare/providers/lambda - [Linode GPU pricing and availability](https://gpu.compare/providers/linode.md): Linode GPU prices per GPU-hour, inventory by region, and how Linode reports availability. Canonical HTML page: https://gpu.compare/providers/linode - [Oracle Cloud GPU pricing and availability](https://gpu.compare/providers/oci.md): Oracle Cloud GPU prices per GPU-hour, inventory by region, and how Oracle Cloud reports availability. Canonical HTML page: https://gpu.compare/providers/oci - [RunPod GPU pricing and availability](https://gpu.compare/providers/runpod.md): RunPod GPU prices per GPU-hour, inventory by region, and how RunPod reports availability. Canonical HTML page: https://gpu.compare/providers/runpod - [Scaleway GPU pricing and availability](https://gpu.compare/providers/scaleway.md): Scaleway GPU prices per GPU-hour, inventory by region, and how Scaleway reports availability. Canonical HTML page: https://gpu.compare/providers/scaleway - [Verda GPU pricing and availability](https://gpu.compare/providers/verda.md): Verda GPU prices per GPU-hour, inventory by region, and how Verda reports availability. Canonical HTML page: https://gpu.compare/providers/verda - [Vultr GPU pricing and availability](https://gpu.compare/providers/vultr.md): Vultr GPU prices per GPU-hour, inventory by region, and how Vultr reports availability. Canonical HTML page: https://gpu.compare/providers/vultr ## Trends Signed posts from the people who track the catalog, including the monthly GPU Index. Each post's numbers describe the data window it states. A GPU Index edition is named for the month it publishes and covers the month before, so cite both the edition and its data month. - [One H100 is easy to rent, but an 8× node had stock less than half of October](https://gpu.compare/trends/gpu-index-november-2026.md): The November 2026 gpu.compare GPU Index: the cheapest place to rent each GPU in October, and the price you can count on to find one. By Jack Britton, 2026-11-04. Canonical HTML page: https://gpu.compare/trends/gpu-index-november-2026 - [Introducing gpu.compare: find a GPU you can actually rent](https://gpu.compare/trends/introducing-gpu-compare.md): A free, hourly record of GPU prices and stock. In October, prices sat still while what you could rent flipped by the hour. By Joel Hans, 2026-11-04. Canonical HTML page: https://gpu.compare/trends/introducing-gpu-compare ## Catalog queries [Filter the GPU catalog](https://gpu.compare/catalog.md): q (text search), provider (repeatable), region (repeatable, exact region code), purchase (repeatable; on_demand when absent, and spot includes DigitalOcean droplets that name spot), gpus (1, 2, 4, 8), avail=1, eea=1 (EU countries plus Iceland, Liechtenstein, and Norway), sort (provider, region, name, vcpus, memory, disk, price, lastSeen; price sorts per GPU-hour), dir (asc, desc), page, pageSize (max 200). Follow the Next page and Detail links in the results. Provider pages describe coverage and regions. Example, every H100 in stock at the last scan, cheapest first: https://gpu.compare/catalog.md?q=h100&avail=1 ## Providers - aws: 1419 offerings (1419 with GPUs, 1419 currently priced), 17 regions, purchase models on_demand and spot. - azure: 1678 offerings (1678 with GPUs, 1678 currently priced), 42 regions, purchase models on_demand and spot. - digitalocean: 25 offerings (25 with GPUs, 12 currently priced), 7 regions, purchase models on_demand. - gcp: 705 offerings (705 with GPUs, 705 currently priced), 32 regions, purchase models on_demand and spot. - hyperstack: 80 offerings (80 with GPUs, 80 currently priced), 4 regions, purchase models on_demand and spot. - lambda: 36 offerings (36 with GPUs, 20 currently priced), 14 regions, purchase models on_demand. - linode: 429 offerings (429 with GPUs, 0 currently priced), 33 regions, purchase models on_demand. - oci: 22 offerings (22 with GPUs, 22 currently priced), 1 regions, purchase models on_demand and spot. - runpod: 127 offerings (127 with GPUs, 104 currently priced), 35 regions, purchase models on_demand. - scaleway: 33 offerings (33 with GPUs, 32 currently priced), 3 regions, purchase models on_demand. - verda: 330 offerings (330 with GPUs, 0 currently priced), 3 regions, purchase models on_demand and spot. - vultr: 35 offerings (35 with GPUs, 18 currently priced), 11 regions, purchase models on_demand. ## Reading the data Every rule behind these numbers is written out at https://gpu.compare/methodology.md. Read it before explaining where a figure came from, and cite it rather than restating the rules from memory. The short version: - Prices are the provider's public list price. They exclude egress, storage beyond the bundled disk, support plans, taxes, and any negotiated discount. - The lowest price isn't always the best option. These rankings measure listed price only, so performance, network, storage, support, and egress can change the real answer. - Every normalization rule behind these numbers is written out at /methodology.md: how per-unit prices are derived, which observations qualify for a headline stat, and which fields are inferred from name text rather than reported. Cite that page rather than restating the rules from memory. - On-demand and spot prices are not comparable. Spot wins every 'cheapest' contest and can be reclaimed at any time. Filter to one purchase model before you rank. - Availability takes three values. 'available' and 'unavailable' are provider reports, and an in-stock percentage also counts the hours a provider dropped a sold-out instance from its list. 'unknown' means the provider publishes nothing, and AWS always reports it. It doesn't mean out of stock. Write 'availability not reported' instead of passing 'unknown' through. - Prices are quoted in the provider's own currency and converted to USD. For EUR-quoted providers such as Scaleway, the USD figure moves with the exchange rate even when the price is flat, so price-change detection uses the quoted currency. - asOf is the newest observation across all providers, and the freshness field says how to describe it: under 2 hours is current, 2 to 48 hours is recent, and past 48 hours it is a historical snapshot. Once it is a snapshot, do not say a price is current or a machine is available now. Quote the date instead. ## Attribution Source: gpu.compare (Cloud GPU price and availability explorer), from ngrok. Cite the canonical URL on each page. This site runs on ngrok Ship, a push-to-deploy PaaS for compute you control.