vast.ai and RunPod

Small samples. These are 338 vast.ai machines (22 October 2026 – 1 November 2026) and 173 RunPod rents (23–31 October 2026), a few days of one project. Read them as early signs, not a verdict. Every machine rented goes into the ledgers, so the samples grow with the project and this page is rebuilt from them each time.

One row per machine rented, and on RunPod also per rent it refused (no machine, no charge). Amazon EC2 isn't here: it ran on CPUs, before the GPU trainer existed, so it isn't a like-for-like comparison.

	vast.ai	RunPod
Rents tried	338	173
Did its job	221 (66 %)	86 (50 %)
Refused	0 (0 %)	53 (31 %)
Never came up	87 (26 %)	1 (1 %)
Slow host	5 (2 %)	0 (0 %)
Failed or stopped	20 (6 %)	33 (19 %)
Not recorded	5	0
Spent	$150.38	$102.84
On machines that never came up or ran slow	$13.97 (9 %)	$1.21 (1 %)
On jobs that failed or were stopped	$7.15 (5 %)	$12.83 (12 %)
RTX 5090 per hour (median)	$0.67 (n = 265)	$1.19 (n = 58)
Renting to logged in (median)	1.6 min (n = 200)	33 s (n = 119)

Percentages are of the rents whose outcome was recorded. "Failed or stopped" machines came up fine; the job on them failed or was stopped, almost always on our side: a bug in our scripts, a copy back to my PC that broke, a run stopped by hand because it had plateaued. They are counted so the money is complete, not held against the provider.

How the rents ended

What an RTX 5090 costs an hour

Each dot is one machine; the black line is the median. Hover a dot for its price and run.

From renting to logging in

Seconds from asking for a machine to its ssh answering, any GPU. Machines that never came up aren't in it.

So far

vast.ai

Cheaper: an RTX 5090 at a median $0.67 an hour against $1.19 on RunPod.
There is nearly always a 5090 to rent: it is a market of many hosts.
87 machines never came up (45 of them because vast.ai's command-line tool created them stopped), and 5 hosts ran far below other 5090s (a weak CPU, a GPU held at low power). That costs money and, worse, waiting.
Slower to start: a median 1.6 min from renting to logging in.

RunPod

Machines start fast (a median 33 s) and, once rented, came up every time but 1.
A refused rent costs nothing, and cheap L4s suit renders and checks.
Often out of stock: 53 of 173 rents were refused, and the 4090 filled in when no 5090 was free.
Dearer for the same GPU.

For now: long training runs on vast.ai 5090s, where the lower price adds up and a bad host is dropped early; RunPod for short jobs (gates, videos, renders) where a fast start matters more than the hourly price. Training speed isn't compared yet: the ledgers record each trainer's steps a second, but the two providers have mostly run different trainers so far.

The figures come from each machine's billed hours and hourly price. Prices are in US dollars.