Cloud GPU pricing comparison: 41 other GPU clouds, model by model.
We keep an index of published prices to set our own — 102 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
- 102Providers tracked for the price guarantee
- Average gap
- −72%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 |
|---|---|---|---|---|
| OnlineGPU vs LiumRecorded 15 Sep 2026 | 42 | −43% | RTX 3070 — $0.07 against $0.05 | Compare |
| OnlineGPU vs Vast.aiRecorded 15 Sep 2026 | 39 | −48% | RTX 3070 — $0.07 against $0.05 | Compare |
| OnlineGPU vs RunPodRecorded 15 Sep 2026 | 34 | −65% | RTX 3070 — $0.13 against $0.05 | Compare |
| OnlineGPU vs GPU.aiRecorded 15 Sep 2026 | 26 | −62% | RTX A4000 — $0.08 against $0.05 | Compare |
| OnlineGPU vs RuncrateRecorded 15 Sep 2026 | 21 | −83% | RTX A4000 — $0.17 against $0.05 | Compare |
| OnlineGPU vs SesterceRecorded 15 Sep 2026 | 21 | −79% | RTX A4000 — $0.17 against $0.05 | Compare |
| OnlineGPU vs EmpirioLabs AIRecorded 15 Sep 2026 | 19 | −84% | RTX A4000 — $0.25 against $0.05 | Compare |
| OnlineGPU vs Massed ComputeRecorded 15 Sep 2026 | 16 | −74% | A30 — $0.35 against $0.08 | Compare |
| OnlineGPU vs GPU OutletRecorded 15 Sep 2026 | 16 | −48% | L4 — $0.13 against $0.08 | Compare |
| OnlineGPU vs AtmosComputeRecorded 15 Sep 2026 | 14 | −83% | RTX 4090 — $0.25 against $0.09 | Compare |
| OnlineGPU vs Omega GradientRecorded 15 Sep 2026 | 14 | −78% | RTX 4090 — $0.39 against $0.09 | Compare |
| OnlineGPU vs SaladRecorded 15 Sep 2026 | 14 | −54% | RTX 3070 — $0.10 against $0.05 | Compare |
| OnlineGPU vs AWSRecorded 15 Sep 2026 | 13 | −89% | T4 — $0.53 against $0.07 | Compare |
| OnlineGPU vs ShadeformRecorded 15 Sep 2026 | 13 | −76% | RTX 4090 — $0.40 against $0.09 | Compare |
| OnlineGPU vs SpheronRecorded 15 Sep 2026 | 13 | −70% | RTX A4000 — $0.17 against $0.05 | Compare |
| OnlineGPU vs IO.NETRecorded 15 Sep 2026 | 12 | −75% | RTX A6000 — $0.49 against $0.17 | Compare |
| OnlineGPU vs Akash NetworkRecorded 15 Sep 2026 | 12 | −65% | RTX 3090 — $0.15 against $0.08 | Compare |
| OnlineGPU vs Theta EdgeCloudRecorded 15 Sep 2026 | 12 | −64% | RTX 3070 — $0.08 against $0.05 | Compare |
| OnlineGPU vs Microsoft AzureRecorded 15 Sep 2026 | 11 | −91% | T4 — $0.53 against $0.07 | Compare |
| OnlineGPU vs ModalRecorded 15 Sep 2026 | 11 | −84% | T4 — $0.59 against $0.07 | Compare |
| OnlineGPU vs HyperstackRecorded 15 Sep 2026 | 11 | −64% | RTX A4000 — $0.15 against $0.05 | Compare |
| OnlineGPU vs VerdaRecorded 15 Sep 2026 | 11 | −56% | V100 16GB — $0.17 against $0.08 | Compare |
| OnlineGPU vs Oracle CloudRecorded 15 Sep 2026 | 10 | −90% | A10 — $2.00 against $0.15 | Compare |
| OnlineGPU vs LambdaRecorded 15 Sep 2026 | 10 | −81% | V100 16GB — $0.79 against $0.08 | Compare |
| OnlineGPU vs OblivusRecorded 15 Sep 2026 | 10 | −79% | RTX A6000 — $0.55 against $0.17 | Compare |
| OnlineGPU vs 1LegionRecorded 15 Sep 2026 | 10 | −69% | RTX 4000 Ada — $0.30 against $0.09 | Compare |
| OnlineGPU vs TensorDockRecorded 15 Sep 2026 | 10 | −68% | RTX A4000 — $0.10 against $0.05 | Compare |
| OnlineGPU vs BeamRecorded 15 Sep 2026 | 10 | −65% | RTX 4090 — $0.44 against $0.09 | Compare |
| OnlineGPU vs PaperspaceRecorded 15 Sep 2026 | 9 | −91% | RTX A4000 — $0.76 against $0.05 | Compare |
| OnlineGPU vs OVHcloudRecorded 15 Sep 2026 | 9 | −85% | V100 16GB — $0.77 against $0.08 | Compare |
| OnlineGPU vs CerebriumRecorded 15 Sep 2026 | 9 | −82% | T4 — $0.59 against $0.07 | Compare |
| OnlineGPU vs Google CloudRecorded 15 Sep 2026 | 9 | −82% | T4 — $0.35 against $0.07 | Compare |
| OnlineGPU vs CrusoeRecorded 15 Sep 2026 | 9 | −81% | A40 — $1.10 against $0.09 | Compare |
| OnlineGPU vs KoyebRecorded 15 Sep 2026 | 9 | −76% | L4 — $0.70 against $0.08 | Compare |
| OnlineGPU vs DigitalOceanRecorded 15 Sep 2026 | 9 | −75% | RTX 4000 Ada — $0.76 against $0.09 | Compare |
| OnlineGPU vs SeewebRecorded 15 Sep 2026 | 9 | −68% | L4 — $0.44 against $0.08 | Compare |
| OnlineGPU vs VultrRecorded 15 Sep 2026 | 9 | −60% | A100 SXM 80GB — $1.29 against $0.29 | Compare |
| OnlineGPU vs AceCloudRecorded 15 Sep 2026 | 8 | −81% | L4 — $0.66 against $0.08 | Compare |
| OnlineGPU vs Together AIRecorded 15 Sep 2026 | 8 | −75% | L40 — $1.49 against $0.19 | Compare |
| OnlineGPU vs CoreWeaveRecorded 15 Sep 2026 | 8 | −68% | L40 — $0.78 against $0.19 | Compare |
| OnlineGPU 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 102 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, $30 of credit, billed per minute. That is the only benchmark that counts.
Billed per minute from the moment the instance is reachable. Minimum credit $30, no subscription.