GPU cloud · 28 GPU models tracked
GPU.ai
$0.17–$8.99per GPU-hour, the cheapest row to the dearest
GPU.ai sells capacity on clouds it connects to and bills you itself, per second, at the rate its catalogue shows. That rate is the lowest it has live for the card at that minute across up to eight regions - its own words are "lowest live on-demand rate per GPU, refreshed continuously" - so it is a floor for the card and not the price of one machine you pick, and it can move between one hour and the next. Cards sold in two tiers carry two rates: the plain one, and the cheaper "community" supply the page marks as such.
As of , GPU.ai lists 28 GPU models we track, from $0.17 per GPU-hour (NVIDIA RTX A4000, community, from) to $8.99 (NVIDIA B300, on-demand, from).
| Website | gpu.ai/ |
|---|---|
| Pricing page | gpu.ai/pricing |
| Regions and tiers | No region or tier named in the rates below: one global rate card |
| Models tracked | 28 |
Price by GPU
As of , per GPU-hour: GPU.ai’s on-demand rate for each card it lists, a cheaper market it also sells, and where the on-demand rate stands among the providers we track.
| GPU | GPU.ai on-demand, per GPU-hour | Against the market (median rate card) |
|---|---|---|
| NVIDIA B300 | $8.99 from captured Deploy ↗ | at the median $8.99 · 26 providers 11 of 26 list lower · lowest Deep Infra $5.89 · compare |
| NVIDIA B200 | $7.99 from captured Deploy ↗ | 15% above the median $6.94 · 40 providers 24 of 40 list lower · lowest Hostinger $4.50 from · compare |
| NVIDIA H200 | $5.29 from captured community $3.59 Deploy ↗ | 16% above the median $4.56 · 56 providers 34 of 56 list lower · lowest Dataoorts $2.29 · compare |
| NVIDIA H200 NVL | $3.29 from captured Deploy ↗ | 9% below the median $3.62 · 7 providers the lowest on-demand rate of the 7 |
| NVIDIA H100 NVL | $3.19 from captured community $2.59 Deploy ↗ | 22% below the median $4.07 · 8 providers 1 of 8 list lower · lowest Massed Compute $2.92 · compare |
| NVIDIA H100 SXM | $2.54 from captured Deploy ↗ | 23% below the median $3.29 · 41 providers 4 of 41 list lower · lowest Cudo Compute $1.79 GPU only · compare |
| NVIDIA H100 PCIe | $1.98 from captured Deploy ↗ | 30% below the median $2.82 · 10 providers the lowest on-demand rate of the 10 |
| NVIDIA RTX PRO 6000 Blackwell | $1.69 from captured Deploy ↗ | 23% below the median $2.20 · 42 providers 6 of 42 list lower · lowest Hostinger $0.65 from · compare |
| NVIDIA A100 80GB | $1.23 from captured community $1.19 Deploy ↗ | 38% below the median $1.99 · 63 providers 4 of 63 list lower · lowest Yotta Labs $0.92 · compare |
| NVIDIA RTX 5090 | $1.19 from captured community $0.69 Deploy ↗ | 51% above the median $0.79 · 23 providers 16 of 23 list lower · lowest OpenRelay $0.50 · compare |
| NVIDIA RTX 4090 | $0.89 from captured community $0.34 Deploy ↗ | 85% above the median $0.48 · 17 providers 12 of 17 list lower · lowest OpenRelay $0.35 · compare |
| NVIDIA L40S | $0.80 from captured community $0.79 Deploy ↗ | 44% below the median $1.42 · 40 providers 1 of 40 list lower · lowest CloudRift $0.63 · compare |
| NVIDIA L40 | $0.78 from captured community $0.69 Deploy ↗ | 20% below the median $0.97 · 11 providers the lowest on-demand rate of the 11 |
| NVIDIA RTX PRO 4500 Blackwell | $0.72 from captured Deploy ↗ | 16% below the median $0.86 · 6 providers the lowest on-demand rate of the 6 |
| NVIDIA RTX 6000 Ada | $0.65 from captured Deploy ↗ | 37% below the median $1.03 · 12 providers the lowest on-demand rate of the 12 |
| NVIDIA A40 | $0.59 from captured Deploy ↗ | 8% below the median $0.64 · 8 providers 1 of 8 list lower · lowest CloudGPU $0.39 · compare |
| NVIDIA L4 | $0.59 from captured Deploy ↗ | 26% below the median $0.80 · 23 providers 6 of 23 list lower · lowest Hikube $0.28 GPU only · compare |
| NVIDIA GeForce RTX 3090 | $0.50 from captured community $0.22 Deploy ↗ | 20% above the median $0.42 · 10 providers 6 of 10 list lower · lowest OpenRelay $0.18 · compare |
| NVIDIA RTX A6000 | $0.40 from captured community $0.33 Deploy ↗ | 32% below the median $0.59 · 21 providers 1 of 21 list lower · lowest Thunder Compute $0.35 · compare |
| NVIDIA RTX 4000 Ada | $0.28 from captured community $0.20 Deploy ↗ | 46% below the median $0.52 · 5 providers the lowest on-demand rate of the 5 |
| NVIDIA RTX A5000 | $0.27 from captured Deploy ↗ | 53% below the median $0.58 · 10 providers the lowest on-demand rate of the 10 |
| NVIDIA RTX A4000 | $0.25 from captured community $0.17 Deploy ↗ | at the median $0.25 · 11 providers 5 of 11 list lower · lowest Hyperstack $0.15 · compare |
| NVIDIA RTX 2000 Ada | $0.24 from captured Deploy ↗ | at the median $0.24 · 2 providers the lowest on-demand rate of the 2 |
The same, in words
- GPU.ai lists the NVIDIA B300 from $8.99 per GPU-hour on-demand (from). The on-demand rate is unchanged since Oct 8, 2026 (7-day change: none). Against the market: at the median rate card ($8.99 across 26 providers); 11 of the 26 providers we track list a lower on-demand rate, the lowest Deep Infra at $5.89 (Deep Infra’s rates, GPU.ai vs Deep Infra).
- GPU.ai lists the NVIDIA B200 from $7.99 per GPU-hour on-demand (from). The on-demand rate is unchanged since Oct 8, 2026 (7-day change: none). Against the market: 15% above the median rate card ($6.94 across 40 providers); 24 of the 40 providers we track list a lower on-demand rate, the lowest Hostinger at $4.50 (Hostinger’s rates, GPU.ai vs Hostinger).
- GPU.ai lists the NVIDIA H200 from $5.29 per GPU-hour on-demand (from), $3.59 as community capacity (from). The on-demand rate is unchanged since Oct 13, 2026 (7-day change: none). Against the market: 16% above the median rate card ($4.56 across 56 providers); 34 of the 56 providers we track list a lower on-demand rate, the lowest Dataoorts at $2.29 (Dataoorts’s rates, GPU.ai vs Dataoorts).
- GPU.ai lists the NVIDIA H200 NVL from $3.29 per GPU-hour on-demand (from). The on-demand rate is unchanged since first read, Oct 1, 2026. Against the market: the lowest on-demand rate of the 7 providers we track for this card, 9% below the median rate card ($3.62 across 7 providers).
- GPU.ai lists the NVIDIA H100 NVL from $3.19 per GPU-hour on-demand (from), $2.59 as community capacity (from). The on-demand rate is unchanged since Nov 4, 2026 (7-day change: none). Against the market: 22% below the median rate card ($4.07 across 8 providers); 1 of the 8 providers we track list a lower on-demand rate, the lowest Massed Compute at $2.92 (Massed Compute’s rates, GPU.ai vs Massed Compute).
- GPU.ai lists the NVIDIA H100 SXM from $2.54 per GPU-hour on-demand (from). The on-demand rate is unchanged since Nov 4, 2026 (7-day change: −36.3%). Against the market: 23% below the median rate card ($3.29 across 41 providers); 4 of the 41 providers we track list a lower on-demand rate, the lowest Cudo Compute at $1.79 (Cudo Compute’s rates, GPU.ai vs Cudo Compute).
- GPU.ai lists the NVIDIA H100 PCIe from $1.98 per GPU-hour on-demand (from). The on-demand rate is unchanged since Nov 4, 2026 (7-day change: −20.2%). Against the market: the lowest on-demand rate of the 10 providers we track for this card, 30% below the median rate card ($2.82 across 10 providers).
- GPU.ai lists the NVIDIA RTX PRO 6000 Blackwell from $1.69 per GPU-hour on-demand (from). The on-demand rate is unchanged since Oct 20, 2026 (7-day change: none). Against the market: 23% below the median rate card ($2.20 across 42 providers); 6 of the 42 providers we track list a lower on-demand rate, the lowest Hostinger at $0.65 (Hostinger’s rates, GPU.ai vs Hostinger).
- GPU.ai lists the NVIDIA A100 80GB from $1.23 per GPU-hour on-demand (from), $1.19 as community capacity (from). The on-demand rate is unchanged since Nov 4, 2026 (7-day change: +15.0%). Against the market: 38% below the median rate card ($1.99 across 63 providers); 4 of the 63 providers we track list a lower on-demand rate, the lowest Yotta Labs at $0.92 (Yotta Labs’s rates, GPU.ai vs Yotta Labs).
- GPU.ai lists the NVIDIA RTX 5090 from $1.19 per GPU-hour on-demand (from), $0.69 as community capacity (from). The on-demand rate is unchanged since Oct 8, 2026 (7-day change: none). Against the market: 51% above the median rate card ($0.79 across 23 providers); 16 of the 23 providers we track list a lower on-demand rate, the lowest OpenRelay at $0.50 (OpenRelay’s rates, GPU.ai vs OpenRelay).
- GPU.ai lists the NVIDIA RTX 4090 from $0.89 per GPU-hour on-demand (from), $0.34 as community capacity (from). The on-demand rate is unchanged since Oct 8, 2026 (7-day change: none). Against the market: 85% above the median rate card ($0.48 across 17 providers); 12 of the 17 providers we track list a lower on-demand rate, the lowest OpenRelay at $0.35 (OpenRelay’s rates, GPU.ai vs OpenRelay).
- GPU.ai lists the NVIDIA L40S from $0.80 per GPU-hour on-demand (from), $0.79 as community capacity (from). The on-demand rate is unchanged since Oct 30, 2026 (7-day change: none). Against the market: 44% below the median rate card ($1.42 across 40 providers); 1 of the 40 providers we track list a lower on-demand rate, the lowest CloudRift at $0.63 (CloudRift’s rates, GPU.ai vs CloudRift).
- GPU.ai lists the NVIDIA L40 from $0.78 per GPU-hour on-demand (from), $0.69 as community capacity (from). The on-demand rate is unchanged since Nov 3, 2026 (7-day change: none). Against the market: the lowest on-demand rate of the 11 providers we track for this card, 20% below the median rate card ($0.97 across 11 providers).
- GPU.ai lists the NVIDIA RTX PRO 4500 Blackwell from $0.72 per GPU-hour on-demand (from). The on-demand rate is unchanged since first read, Oct 1, 2026 (7-day change: none). Against the market: the lowest on-demand rate of the 6 providers we track for this card, 16% below the median rate card ($0.86 across 6 providers).
- GPU.ai lists the NVIDIA RTX 6000 Ada from $0.65 per GPU-hour on-demand (from). The on-demand rate is unchanged since Nov 3, 2026 (7-day change: none). Against the market: the lowest on-demand rate of the 12 providers we track for this card, 37% below the median rate card ($1.03 across 12 providers).
- GPU.ai lists the NVIDIA A40 from $0.59 per GPU-hour on-demand (from). The on-demand rate is unchanged since Oct 8, 2026 (7-day change: none). Against the market: 8% below the median rate card ($0.64 across 8 providers); 1 of the 8 providers we track list a lower on-demand rate, the lowest CloudGPU at $0.39 (CloudGPU’s rates, GPU.ai vs CloudGPU).
- GPU.ai lists the NVIDIA L4 from $0.59 per GPU-hour on-demand (from). The on-demand rate is unchanged since Oct 8, 2026 (7-day change: none). Against the market: 26% below the median rate card ($0.80 across 23 providers); 6 of the 23 providers we track list a lower on-demand rate, the lowest Hikube at $0.28 (Hikube’s rates, GPU.ai vs Hikube).
- GPU.ai lists the NVIDIA GeForce RTX 3090 from $0.50 per GPU-hour on-demand (from), $0.22 as community capacity (from). The on-demand rate is unchanged since first read, Oct 1, 2026 (7-day change: none). Against the market: 20% above the median rate card ($0.42 across 10 providers); 6 of the 10 providers we track list a lower on-demand rate, the lowest OpenRelay at $0.18 (OpenRelay’s rates, GPU.ai vs OpenRelay).
- GPU.ai lists the NVIDIA RTX A6000 from $0.40 per GPU-hour on-demand (from), $0.33 as community capacity (from). The on-demand rate is unchanged since Nov 4, 2026 (7-day change: none). Against the market: 32% below the median rate card ($0.59 across 21 providers); 1 of the 21 providers we track list a lower on-demand rate, the lowest Thunder Compute at $0.35 (Thunder Compute’s rates, GPU.ai vs Thunder Compute).
- GPU.ai lists the NVIDIA RTX 4000 Ada from $0.28 per GPU-hour on-demand (from), $0.20 as community capacity (from). The on-demand rate is unchanged since first read, Oct 2, 2026 (7-day change: none). Against the market: the lowest on-demand rate of the 5 providers we track for this card, 46% below the median rate card ($0.52 across 5 providers).
- GPU.ai lists the NVIDIA RTX A5000 from $0.27 per GPU-hour on-demand (from). The on-demand rate is unchanged since first read, Oct 1, 2026 (7-day change: none). Against the market: the lowest on-demand rate of the 10 providers we track for this card, 53% below the median rate card ($0.58 across 10 providers).
- GPU.ai lists the NVIDIA RTX A4000 from $0.25 per GPU-hour on-demand (from), $0.17 as community capacity (from). The on-demand rate is unchanged since Nov 5, 2026 (7-day change: none). Against the market: at the median rate card ($0.25 across 11 providers); 5 of the 11 providers we track list a lower on-demand rate, the lowest Hyperstack at $0.15 (Hyperstack’s rates, GPU.ai vs Hyperstack).
- GPU.ai lists the NVIDIA RTX 2000 Ada from $0.24 per GPU-hour on-demand (from). The on-demand rate is unchanged since first read, Oct 2, 2026 (7-day change: none). Against the market: the lowest on-demand rate of the 2 providers we track for this card, at the median rate card ($0.24 across 2 providers).
All listed prices
One table per price type, never mixed. Every rate normalized to one GPU for one hour; a price opens the page it was read from, and 7d marks a rate that moved in the past seven days. The day each figure was first listed is in this page’s data, below.
On-demand
Ask an AI agent about GPU.ai’s rates: Claude, Cursor, VS Code and ChatGPT can read them through the GPU Economy MCP server - free, no key, every answer citing the page and read time.
Every rate on this page as data: /providers/gpuai/prices.json — the same fields as prices.json, narrowed to this provider. CC BY 4.0, no key, no rate limit. On your own page: embed this table (/embed/providers/gpuai, an iframe, no script).
Compare GPU.ai with
Every provider that lists at least two of the same GPUs on-demand, most shared GPUs first; each opens the side-by-side page.
- RunPod 23 shared GPUs
- EmpirioLabs 19 shared GPUs
- Vast.ai 19 shared GPUs
- Lium 15 shared GPUs
- Sesterce 15 shared GPUs
- HexGrid 12 shared GPUs
- Massed Compute 11 shared GPUs
- Hyperstack 10 shared GPUs
- Amazon Web Services 9 shared GPUs
- Beam 9 shared GPUs
- Paperspace 9 shared GPUs
- Verda 9 shared GPUs
- Yotta Labs 9 shared GPUs
- Barrack AI 8 shared GPUs
- Modal 8 shared GPUs
- OpenRelay 8 shared GPUs
- CloudGPU 7 shared GPUs
- CoreWeave 7 shared GPUs
- E2E Networks 7 shared GPUs
- Koyeb 7 shared GPUs
- Oracle Cloud (OCI) 7 shared GPUs
- QuantaCloud 7 shared GPUs
- Seeweb 7 shared GPUs
- Cerebrium 6 shared GPUs
- Daytona 6 shared GPUs
- Geodd 6 shared GPUs
- Hostinger 6 shared GPUs
- Hugging Face 6 shared GPUs
- immers.cloud 6 shared GPUs
- Nebius 6 shared GPUs
- Trooper.AI 6 shared GPUs
- Airon 5 shared GPUs
- Civo 5 shared GPUs
- Crusoe 5 shared GPUs
- Dataoorts 5 shared GPUs
- Gcore 5 shared GPUs
- Google Cloud 5 shared GPUs
- Hikube 5 shared GPUs
- Jarvislabs 5 shared GPUs
- Lambda 5 shared GPUs
- Limestone Networks 5 shared GPUs
- Lyceum 5 shared GPUs
- Microsoft Azure 5 shared GPUs
- Nebula Block 5 shared GPUs
- Northflank 5 shared GPUs
- OVHcloud 5 shared GPUs
- packet.ai 5 shared GPUs
- Runware 5 shared GPUs
- SwissGPU 5 shared GPUs
- SynpixCloud 5 shared GPUs
- Tata Communications 5 shared GPUs
- airgpu 4 shared GPUs
- Clore.ai 4 shared GPUs
- CloudRift 4 shared GPUs
- Deep Infra 4 shared GPUs
- fal 4 shared GPUs
- FarmGPU 4 shared GPUs
- FPT AI Factory 4 shared GPUs
- HPC-AI.COM 4 shared GPUs
- Lightning AI 4 shared GPUs
- Neysa 4 shared GPUs
- Octa.space 4 shared GPUs
- Servercore 4 shared GPUs
- Thunder Compute 4 shared GPUs
- Together AI 4 shared GPUs
- Baseten 3 shared GPUs
- Bitdeer AI 3 shared GPUs
- centron 3 shared GPUs
- Cloud.ru Evolution 3 shared GPUs
- Cudo Compute 3 shared GPUs
- Denvr Dataworks 3 shared GPUs
- Elice Cloud 3 shared GPUs
- Fireworks AI 3 shared GPUs
- Hyperbolic 3 shared GPUs
- Intelion Cloud 3 shared GPUs
- Nova Cloud 3 shared GPUs
- RunComfy 3 shared GPUs
- Salad 3 shared GPUs
- Selectel 3 shared GPUs
- STACKIT 3 shared GPUs
- VESSL AI 3 shared GPUs
- Atlantic.Net 2 shared GPUs
- Atlas Cloud 2 shared GPUs
- Canopy Wave 2 shared GPUs
- Enverge 2 shared GPUs
- FlexAI 2 shared GPUs
- FluidCore 2 shared GPUs
- GMI Cloud 2 shared GPUs
- Infomaniak 2 shared GPUs
- Leafcloud 2 shared GPUs
- Replicate 2 shared GPUs
- Taiga Cloud (Northern Data Group) 2 shared GPUs
Data updated: