GPU.ai GPU pricing
21 GPUs priced today across 34 configurations. On average 9.5% below the market median.
Prices collected 2026-09-18 · seen since 2026-09-11
Visit GPU.ai →GPU.ai's on-demand median per GPU-hour, per snapshot — its eight most expensive GPUs. Gaps are runs where the GPU was not listed.
At a glance
Every GPU GPU.ai prices, against the market
Per GPU-hour, on-demand, 2026-09-18. vs market compares GPU.ai's own median for that GPU with the index median across every provider. Click a GPU for every provider's price.
| GPU | GPU.ai | Market | vs market | Per month | Sizes | Billing |
|---|---|---|---|---|---|---|
 NVIDIA B200 | $7.78 | $6.59 | +18% | $5,598 | 1x · 2 configs | on-demand |
| NVIDIA H100 SXM | $3.69 | $3.50 | +5% | $2,653 | 1x · 2 configs | on-demand |
| NVIDIA H200 | $3.59 | $4.45 | -19% | $2,585 | 1x · 1 config | on-demand |
| NVIDIA H200 NVL | $3.46 | $3.72 | -7% | $2,488 | 1x · 2 configs | on-demand |
| NVIDIA H100 NVL | $3.05 | $3.03 | +1% | $2,196 | 1x · 2 configs | on-demand |
| NVIDIA H100 PCIe | $2.43 | $2.88 | -16% | $1,750 | 1x · 2 configs | on-demand |
| NVIDIA RTX PRO 6000 Blackwell | $1.84 | $2.05 | -10% | $1,325 | 1x · 2 configs | on-demand |
| NVIDIA A100 80GB | $1.21 | $1.79 | -32% | $871 | 1x · 2 configs | on-demand |
| NVIDIA A100 40GB | $0.89 | $1.48 | -40% | $641 | 1x · 1 config | on-demand |
| NVIDIA L40S | $0.80 | $1.47 | -46% | $572 | 1x · 2 configs | on-demand |
| NVIDIA L40 | $0.78 | $0.86 | -9% | $562 | 1x · 1 config | on-demand |
| NVIDIA RTX 5090cheapest anywhere | $0.73 | $0.69 | +6% | $526 | 1x · 2 configs | on-demand |
| NVIDIA RTX 5080 | $0.67 | $0.53 | +25% | $482 | 1x · 1 config | on-demand |
| NVIDIA RTX 6000 Ada | $0.65 | $0.82 | -20% | $468 | 1x · 1 config | on-demand |
| NVIDIA RTX 4090 | $0.60 | $0.60 | 0% | $432 | 1x · 2 configs | on-demand |
| NVIDIA A40 | $0.49 | $0.65 | -25% | $353 | 1x · 1 config | on-demand |
| NVIDIA RTX A6000 | $0.44 | $0.56 | -22% | $313 | 1x · 2 configs | on-demand |
| NVIDIA L4cheapest anywhere | $0.38 | $0.80 | -52% | $274 | 1x · 2 configs | on-demand |
| NVIDIA RTX 4080 | $0.34 | $0.31 | +11% | $245 | 1x · 1 config | on-demand |
| NVIDIA RTX 3090 | $0.34 | $0.25 | +34% | $241 | 1x · 2 configs | on-demand |
| NVIDIA RTX A5000 | $0.27 | $0.27 | 0% | $194 | 1x · 1 config | on-demand |
Price history
GPU.ai has been in the index since 2026-09-11. The chart at the top draws its on-demand median per GPU on every snapshot; a gap is a run where the GPU was not listed, never a guess.
Recent price changes
| When | GPU | From | To | Change |
|---|---|---|---|---|
| 2026-09-18 22:13 | NVIDIA RTX 4090 | $0.54 | $0.60 | +11.1% |
| 2026-09-18 22:13 | NVIDIA A100 80GB | $1.10 | $1.21 | +10.5% |
| 2026-09-18 22:13 | NVIDIA B200 | $7.09 | $7.78 | +9.7% |
| 2026-09-18 22:13 | NVIDIA RTX A5000 | $0.26 | $0.27 | +5.9% |
| 2026-09-18 22:13 | NVIDIA RTX 3090 | $0.32 | $0.34 | +4.7% |
| 2026-09-18 22:13 | NVIDIA H200 NVL | $3.62 | $3.46 | -4.4% |
| 2026-09-18 22:13 | NVIDIA H100 NVL | $3.19 | $3.05 | -4.4% |
| 2026-09-18 20:56 | NVIDIA RTX 5080 | $0.31 | $0.67 | +116.1% |
| 2026-09-18 20:56 | NVIDIA RTX 3090 | $0.15 | $0.32 | +113.3% |
| 2026-09-18 20:56 | NVIDIA RTX 5090 | $0.46 | $0.74 | +59.8% |
| 2026-09-18 20:56 | NVIDIA RTX 4090 | $0.41 | $0.54 | +31.7% |
| 2026-09-18 20:56 | NVIDIA H100 PCIe | $1.99 | $2.43 | +22.1% |
A change is the provider's median for that GPU moving between two consecutive snapshots. Listings appearing or disappearing are not counted.
Compare GPU.ai
Other marketplaces
RunPod Community Cloud25 GPUs · -30.8% vs market
Vast.ai23 GPUs · +0.1% vs market
io.net21 GPUs · -7.1% vs marketQuestions
How much does GPU.ai charge for an H100?
As of 2026-09-18, GPU.ai's on-demand H100 SXM works out to $3.69 per GPU-hour, 5% above the market median of $3.50. That is about $2,653 a month per GPU at 720 hours.
Is GPU.ai cheap?
Across the 21 GPUs it prices, GPU.ai averages 9.5% below the market median, and is cheaper than the median on 12 of them. It is the single cheapest provider we track for RTX 5090, L4. As a marketplace its listings are independent hosts, so the price is a spread and the cheapest listing may be a single machine.
Where do these GPU.ai prices come from?
Public pricing page, parsed. Per-GPU price is the listed per-GPU rate, or the node price divided by GPU count. Prices near reserved, committed or contract terms are discarded; anything outside a plausible range for the GPU is discarded. Prices are collected every six hours and every snapshot is kept, so the history above is what GPU.ai actually published on those dates. Nothing is estimated or interpolated.
Does GPU.ai offer spot or reserved pricing?
No spot or interruptible rates were listed on the pricing we read. Reserved pricing, where it exists, is by quote and not published. Only on-demand prices enter the index median.