Compared with 41 other GPU clouds, model by model.
We keep an index of published prices to set our own — 137 providers, recorded September 2026. These pages show what that index says about the providers people actually ask us about, with the source on every row.
- Providers compared
- 41With at least 10 GPU models in common
- Index size
- 137Providers tracked for the price guarantee
- Average gap
- −73%Our on-demand rate against theirs
- Recorded
- September 2026Prices move; each row links to its source
Pick the provider you are pricing against.
Ranked by how many GPU models we have in common — the more overlap, the more the average gap means something.
| Provider | Models compared | Average gap | Cheapest overlap | Open |
|---|---|---|---|---|
| SpotGPUs vs LiumRecorded 15 Sep 2026 | 42 | −43% | RTX 3070 — $0.07 against $0.05 | Compare |
| SpotGPUs vs Vast.aiRecorded 15 Sep 2026 | 39 | −48% | RTX 3070 — $0.07 against $0.05 | Compare |
| SpotGPUs vs RunPodRecorded 15 Sep 2026 | 34 | −65% | RTX 3070 — $0.13 against $0.05 | Compare |
| SpotGPUs vs GPU.aiRecorded 15 Sep 2026 | 26 | −62% | RTX A4000 — $0.08 against $0.05 | Compare |
| SpotGPUs vs RuncrateRecorded 15 Sep 2026 | 21 | −83% | RTX A4000 — $0.17 against $0.05 | Compare |
| SpotGPUs vs SesterceRecorded 15 Sep 2026 | 21 | −79% | RTX A4000 — $0.17 against $0.05 | Compare |
| SpotGPUs vs EmpirioLabs AIRecorded 15 Sep 2026 | 19 | −84% | RTX A4000 — $0.25 against $0.05 | Compare |
| SpotGPUs vs Massed ComputeRecorded 15 Sep 2026 | 16 | −74% | A30 — $0.35 against $0.08 | Compare |
| SpotGPUs vs GPU OutletRecorded 15 Sep 2026 | 16 | −48% | L4 — $0.13 against $0.08 | Compare |
| SpotGPUs vs AtmosComputeRecorded 15 Sep 2026 | 14 | −83% | RTX 4090 — $0.25 against $0.09 | Compare |
| SpotGPUs vs Omega GradientRecorded 15 Sep 2026 | 14 | −78% | RTX 4090 — $0.39 against $0.09 | Compare |
| SpotGPUs vs SaladRecorded 15 Sep 2026 | 14 | −59% | RTX 3070 — $0.10 against $0.05 | Compare |
| SpotGPUs vs AWSRecorded 15 Sep 2026 | 13 | −89% | T4 — $0.53 against $0.07 | Compare |
| SpotGPUs vs ShadeformRecorded 15 Sep 2026 | 13 | −76% | RTX 4090 — $0.40 against $0.09 | Compare |
| SpotGPUs vs SpheronRecorded 15 Sep 2026 | 13 | −70% | RTX A4000 — $0.17 against $0.05 | Compare |
| SpotGPUs vs IO.NETRecorded 15 Sep 2026 | 12 | −75% | RTX A6000 — $0.49 against $0.17 | Compare |
| SpotGPUs vs Akash NetworkRecorded 15 Sep 2026 | 12 | −65% | RTX 3090 — $0.15 against $0.08 | Compare |
| SpotGPUs vs Theta EdgeCloudRecorded 15 Sep 2026 | 12 | −64% | RTX 3070 — $0.08 against $0.05 | Compare |
| SpotGPUs vs Microsoft AzureRecorded 15 Sep 2026 | 11 | −91% | T4 — $0.53 against $0.07 | Compare |
| SpotGPUs vs ModalRecorded 15 Sep 2026 | 11 | −84% | T4 — $0.59 against $0.07 | Compare |
| SpotGPUs vs HyperstackRecorded 15 Sep 2026 | 11 | −64% | RTX A4000 — $0.15 against $0.05 | Compare |
| SpotGPUs vs VerdaRecorded 15 Sep 2026 | 11 | −56% | V100 16GB — $0.17 against $0.08 | Compare |
| SpotGPUs vs Oracle CloudRecorded 15 Sep 2026 | 10 | −90% | A10 — $2.00 against $0.15 | Compare |
| SpotGPUs vs LambdaRecorded 15 Sep 2026 | 10 | −81% | V100 16GB — $0.79 against $0.08 | Compare |
| SpotGPUs vs OblivusRecorded 15 Sep 2026 | 10 | −79% | RTX A6000 — $0.55 against $0.17 | Compare |
| SpotGPUs vs 1LegionRecorded 15 Sep 2026 | 10 | −69% | RTX 4000 Ada — $0.30 against $0.09 | Compare |
| SpotGPUs vs TensorDockRecorded 15 Sep 2026 | 10 | −68% | RTX A4000 — $0.10 against $0.05 | Compare |
| SpotGPUs vs BeamRecorded 15 Sep 2026 | 10 | −65% | RTX 4090 — $0.44 against $0.09 | Compare |
| SpotGPUs vs PaperspaceRecorded 15 Sep 2026 | 9 | −91% | RTX A4000 — $0.76 against $0.05 | Compare |
| SpotGPUs vs OVHcloudRecorded 15 Sep 2026 | 9 | −85% | V100 16GB — $0.77 against $0.08 | Compare |
| SpotGPUs vs CerebriumRecorded 15 Sep 2026 | 9 | −82% | T4 — $0.59 against $0.07 | Compare |
| SpotGPUs vs Google CloudRecorded 15 Sep 2026 | 9 | −82% | T4 — $0.35 against $0.07 | Compare |
| SpotGPUs vs CrusoeRecorded 15 Sep 2026 | 9 | −81% | A40 — $1.10 against $0.09 | Compare |
| SpotGPUs vs KoyebRecorded 15 Sep 2026 | 9 | −76% | L4 — $0.70 against $0.08 | Compare |
| SpotGPUs vs DigitalOceanRecorded 15 Sep 2026 | 9 | −75% | RTX 4000 Ada — $0.76 against $0.09 | Compare |
| SpotGPUs vs SeewebRecorded 15 Sep 2026 | 9 | −68% | L4 — $0.44 against $0.08 | Compare |
| SpotGPUs vs VultrRecorded 15 Sep 2026 | 9 | −60% | A100 SXM 80GB — $1.29 against $0.29 | Compare |
| SpotGPUs vs AceCloudRecorded 15 Sep 2026 | 8 | −81% | L4 — $0.66 against $0.08 | Compare |
| SpotGPUs vs Together AIRecorded 15 Sep 2026 | 8 | −75% | L40 — $1.49 against $0.19 | Compare |
| SpotGPUs vs CoreWeaveRecorded 15 Sep 2026 | 8 | −68% | L40 — $0.78 against $0.19 | Compare |
| SpotGPUs vs Novita AIRecorded 15 Sep 2026 | 8 | −61% | RTX 4090 — $0.33 against $0.09 | Compare |
Provider names are trademarks of their owners. These pages compare published prices on a stated date; they are not affiliated with, endorsed by, or sourced from the providers themselves unless the row links to their own page.
A price claim nobody can check is worth nothing.
The index sets our prices
Every tier of every model is priced at or below 80% of the lowest published rate in this index. The comparison pages are simply that index, made readable.
Dated and sourced
Every figure carries the date it was recorded and a link to where it came from. Prices move — an undated comparison is a marketing claim, not a fact.
Price only
We compare the one thing we can verify from outside. Capacity, support and interruption behaviour differ between providers, and we do not pretend to measure them.
About these comparisons
Where do these prices come from?
From the price index we build to set our own catalogue: 2,158 published offers from 137 providers, recorded September 2026. Each row on a comparison page links to the page the figure came from.
Why would a provider be missing?
We only build a page when a provider publishes prices for at least ten GPU models we also run — below that, a comparison says more about the gaps than about the prices. The full index still feeds our price guarantee.
Do you compare anything other than price?
No. Capacity, support, network and interruption behaviour are not things we can measure on someone else's platform, so we do not assert them. What we do promise on our own side is written on the pricing page and in the SLA.
Compare the invoice, not the table.
One job, $40 of credit, billed per minute. That is the only benchmark that counts.
Billed per minute from the moment the instance is reachable. Minimum credit $40, no subscription.