Category guide
Compare GPU Cloud: category guide
Answer in brief
Runcrate is our editor’s choice for comparing GPU clouds: the best developer platform we know, billing by the second, and GPUs you can start now, in about four minutes.
A GPU cloud comparison works when you apply the criteria in a fixed order: first whether you can start the GPU today, then the smallest amount you can rent, then the billing unit, and only then the price per GPU-hour. The order matters because each step removes options the next step would have wasted your time on.
Compare in order, not all at once
Most comparisons of GPU clouds start with a price column. That is the wrong place to start. A price says nothing about whether the machine exists for you, how many cards you must take, or how your time is rounded before the rate is applied.
So this site compares in a fixed order. Availability comes first, because a GPU you cannot start has no price. The smallest rental comes second, because it decides what an hour costs. The billing unit comes third. Price is fourth, and by then only a few options are left to read.
What a fair GPU cloud comparison holds constant
Two prices compare only when they are for the same thing: the same GPU, the same pricing model, the same number of cards and the same day. Change any one and you are comparing two products, not two providers.
Every table here holds those four constant. The prices are on demand, per GPU-hour, for one named GPU, read on one date. On 12 Nov 2026, 10 providers listed an on-demand B200 price, from $3.50 to $14.24 per GPU-hour.
The order to compare GPU clouds in
Work through the questions in this order. Each one removes options the next would have wasted time on.
- Can you start the GPU today?
Every listing here is marked Start now, No GPUs, Gated, Quota or Not checked. Drop what you cannot start before you read a price.
- What is the smallest thing you can rent?
One card and an 8-GPU node are different purchases. Compare within one size only.
- What unit is your time billed in?
Seconds, minutes or hours. Ask too whether the minutes a machine spends booting are billed.
- Is the price like for like?
Same GPU, on demand, per GPU-hour, read on the same day. A spot or reserved figure belongs in a separate column.
- What else is on the bill?
Storage that outlives the machine and data leaving the provider. Neither is in the GPU-hour.
- How will you drive it?
A console is enough for one machine. For repeated work you need an API, a CLI or a tool an agent can use.
Terms this guide uses
- Like for like
A like-for-like comparison sets two prices side by side only when the GPU, the pricing model, the number of cards and the date are the same.
- On-demand
An on-demand price is what you pay to start a machine when you want it and stop it when you are done, with no term and no risk of losing it.
- Smallest rental
The smallest rental is the least you can rent from a listing: one GPU, or a whole machine of several, most often eight.
- Billing unit
The billing unit is the step your running time is rounded up to before the hourly rate is applied: a second, a minute or an hour.
- Availability label
An availability label is this site’s dated record of whether a listed GPU can be started: Start now, No GPUs, Gated, Quota or Not checked.
Where Runcrate fits
Compare GPU Cloud names one provider as its editor’s choice and gives the reasons in full, so you can weigh them against your own brief.
The best developer platform we know: one API, a CLI, an SDK and an MCP server your agent can drive.
Billed by the second on most GPUs. Booting is never billed.
GPUs you can start now, in about four minutes.
26 GPU models from $0.23/hr, on one prepaid balance.
Questions and answers
Sources for this entry
- Runcrate: Runcrate’s official site: what the platform is, how it bills, its developer tools and the range of GPUs it offers. Read 12 Nov 2026.
- Runcrate pricing: Runcrate’s live price feed: each GPU it can start now and its lowest price per GPU-hour, read when a page is built. Read 12 Nov 2026.
- Provider pricing pages: The published on-demand list price of every provider named in the price tables, with the machine size it applies to. Read 12 Nov 2026.
- Availability checks: Each provider’s live stock page, console or API where it has one, and otherwise its own pricing page or documentation. Read 12 Nov 2026.
- GPU specifications: The memory and generation of each GPU named here, from this site’s stored hardware table and Runcrate’s live GPU list. Read 12 Nov 2026.
Checklists are original editorial guidance. Published prices do not establish availability.
Continue this reading path
Cite this entry
Compare GPU Cloud. “Compare GPU Cloud: category guide.” Content reviewed 12 Nov 2026. https://comparegpucloud.com/guide