Cheapest cloud GPU right now: price per GPU-hour by chip and region
Lowest on-demand list price per GPU-hour on 2026-10-13 18:10 UTC: H100 USD 6.8800 per GPU-hour (AWS us-east-1); A100 80GB USD 3.4309 per GPU-hour (AWS us-east-1); T4 USD 0.5260 per GPU-hour (AWS us-east-1). Every chip and region is listed separately below.
Page generated 2026-10-13 18:10 UTC, from the providers' public price lists
H100
GPU memory per GPU: 80 to 94 GB. Cheapest on-demand first.
| Provider | Region | Instance | GPUs per instance | On-demand USD per GPU-hour | Spot USD per GPU-hour | Price date | Source |
|---|---|---|---|---|---|---|---|
| AWS | us-east-1 | p5.48xlarge | 8 | 6.8800 | not in this version | 2026-10-12 | AWS Price List |
| AWS | us-west-2 | p5.48xlarge | 8 | 6.8800 | not in this version | 2026-10-12 | AWS Price List |
| Azure | eastus | NC40ads_H100_v5 | 1 | 6.9800 | 1.2899 | 2026-08-01 | Azure Retail Prices API |
| Azure | westus2 | NC40ads_H100_v5 | 1 | 6.9800 | 3.2164 | 2026-10-01 | Azure Retail Prices API |
| Azure | southcentralus | NC40ads_H100_v5 | 1 | 8.3800 | 2.1394 | 2026-10-01 | Azure Retail Prices API |
| Azure | southeastasia | NC40ads_H100_v5 | 1 | 9.0740 | 1.6769 | 2026-09-01 | Azure Retail Prices API |
| Azure | westeurope | NC40ads_H100_v5 | 1 | 9.0800 | 3.3469 | 2026-10-01 | Azure Retail Prices API |
A100 80GB
GPU memory per GPU: 80 GB. Cheapest on-demand first.
| Provider | Region | Instance | GPUs per instance | On-demand USD per GPU-hour | Spot USD per GPU-hour | Price date | Source |
|---|---|---|---|---|---|---|---|
| AWS | us-east-1 | p4de.24xlarge | 8 | 3.4309 | not in this version | 2026-10-12 | AWS Price List |
| AWS | us-west-2 | p4de.24xlarge | 8 | 3.4309 | not in this version | 2026-10-12 | AWS Price List |
| Azure | eastus | NC24ads_A100_v4 | 1 | 3.6730 | 0.6788 | 2026-07-01 | Azure Retail Prices API |
| Azure | westus2 | NC24ads_A100_v4 | 1 | 3.6730 | 0.6788 | 2026-08-01 | Azure Retail Prices API |
| AWS | ap-southeast-1 | p4de.24xlarge | 8 | 4.1171 | not in this version | 2026-10-12 | AWS Price List |
| Azure | southcentralus | NC48ads_A100_v4 | 2 | 4.4075 | 0.8145 | 2026-04-01 | Azure Retail Prices API |
| Azure | southeastasia | NC24ads_A100_v4 | 1 | 4.7750 | 0.8824 | 2025-11-01 | Azure Retail Prices API |
| Azure | westeurope | NC24ads_A100_v4 | 1 | 4.7750 | 0.8824 | 2026-09-01 | Azure Retail Prices API |
A100 40GB
GPU memory per GPU: 40 GB. Cheapest on-demand first.
| Provider | Region | Instance | GPUs per instance | On-demand USD per GPU-hour | Spot USD per GPU-hour | Price date | Source |
|---|---|---|---|---|---|---|---|
| AWS | us-east-1 | p4d.24xlarge | 8 | 2.7447 | not in this version | 2026-10-12 | AWS Price List |
| AWS | us-west-2 | p4d.24xlarge | 8 | 2.7447 | not in this version | 2026-10-12 | AWS Price List |
| AWS | eu-west-1 | p4d.24xlarge | 8 | 2.9645 | not in this version | 2026-10-12 | AWS Price List |
| AWS | ap-southeast-1 | p4d.24xlarge | 8 | 3.2936 | not in this version | 2026-10-12 | AWS Price List |
| Azure | eastus | ND96asr_A100_v4 | 8 | 3.3996 | 0.7479 | 2025-03-01 | Azure Retail Prices API |
| Azure | westus2 | ND96asr_A100_v4 | 8 | 3.3996 | 0.7479 | 2025-02-01 | Azure Retail Prices API |
| Azure | southcentralus | ND96asr_A100_v4 | 8 | 4.0788 | 0.8998 | 2025-02-01 | Azure Retail Prices API |
| Azure | westeurope | ND96asr_A100_v4 | 8 | 4.4196 | 0.9723 | 2025-02-01 | Azure Retail Prices API |
L40S
GPU memory per GPU: 48 GB. Cheapest on-demand first.
| Provider | Region | Instance | GPUs per instance | On-demand USD per GPU-hour | Spot USD per GPU-hour | Price date | Source |
|---|---|---|---|---|---|---|---|
| AWS | us-east-1 | g6e.xlarge | 1 | 1.8610 | not in this version | 2026-10-12 | AWS Price List |
| AWS | us-west-2 | g6e.xlarge | 1 | 1.8610 | not in this version | 2026-10-12 | AWS Price List |
L4
GPU memory per GPU: 24 GB. Cheapest on-demand first.
| Provider | Region | Instance | GPUs per instance | On-demand USD per GPU-hour | Spot USD per GPU-hour | Price date | Source |
|---|---|---|---|---|---|---|---|
| AWS | us-east-1 | g6.xlarge | 1 | 0.8048 | not in this version | 2026-10-12 | AWS Price List |
| AWS | us-west-2 | g6.xlarge | 1 | 0.8048 | not in this version | 2026-10-12 | AWS Price List |
A10
GPU memory per GPU: 24 GB. Cheapest on-demand first.
| Provider | Region | Instance | GPUs per instance | On-demand USD per GPU-hour | Spot USD per GPU-hour | Price date | Source |
|---|---|---|---|---|---|---|---|
| Azure | eastus | NV36ads_A10_v5 | 1 | 3.2000 | 0.5914 | 2025-12-01 | Azure Retail Prices API |
| Azure | westus2 | NV36ads_A10_v5 | 1 | 3.2000 | 0.5914 | 2026-05-01 | Azure Retail Prices API |
| Azure | southcentralus | NV36ads_A10_v5 | 1 | 3.8400 | 0.7096 | 2025-09-01 | Azure Retail Prices API |
| Azure | southeastasia | NV36ads_A10_v5 | 1 | 4.1600 | 0.7688 | 2025-09-01 | Azure Retail Prices API |
| Azure | westeurope | NV36ads_A10_v5 | 1 | 4.1600 | 0.7688 | 2026-06-01 | Azure Retail Prices API |
A10G
GPU memory per GPU: 24 GB. Cheapest on-demand first.
| Provider | Region | Instance | GPUs per instance | On-demand USD per GPU-hour | Spot USD per GPU-hour | Price date | Source |
|---|---|---|---|---|---|---|---|
| AWS | us-east-1 | g5.xlarge | 1 | 1.0060 | not in this version | 2026-10-12 | AWS Price List |
| AWS | us-west-2 | g5.xlarge | 1 | 1.0060 | not in this version | 2026-10-12 | AWS Price List |
| AWS | eu-west-1 | g5.xlarge | 1 | 1.1230 | not in this version | 2026-10-12 | AWS Price List |
T4
GPU memory per GPU: 16 GB. Cheapest on-demand first.
| Provider | Region | Instance | GPUs per instance | On-demand USD per GPU-hour | Spot USD per GPU-hour | Price date | Source |
|---|---|---|---|---|---|---|---|
| AWS | us-east-1 | g4dn.xlarge | 1 | 0.5260 | not in this version | 2026-10-12 | AWS Price List |
| AWS | us-west-2 | g4dn.xlarge | 1 | 0.5260 | not in this version | 2026-10-12 | AWS Price List |
| Azure | eastus | NC4as_T4_v3 | 1 | 0.5260 | 0.1492 | 2026-09-01 | Azure Retail Prices API |
| Azure | westus2 | NC4as_T4_v3 | 1 | 0.5260 | 0.1605 | 2026-10-01 | Azure Retail Prices API |
| AWS | eu-west-1 | g4dn.xlarge | 1 | 0.5870 | not in this version | 2026-10-12 | AWS Price List |
| Azure | southcentralus | NC4as_T4_v3 | 1 | 0.6310 | 0.1790 | 2026-09-01 | Azure Retail Prices API |
| Azure | westeurope | NC4as_T4_v3 | 1 | 0.6580 | 0.2925 | 2026-10-01 | Azure Retail Prices API |
| AWS | ap-southeast-1 | g4dn.xlarge | 1 | 0.7360 | not in this version | 2026-10-12 | AWS Price List |
| Azure | southeastasia | NC4as_T4_v3 | 1 | 0.7360 | 0.2087 | 2026-09-01 | Azure Retail Prices API |
How this is counted
- Official list prices only, taken from the Azure Retail Prices API and the AWS Price List files. Negotiated discounts, savings plans, reservations and credits are not included.
- Not in this version: Google Cloud, AWS spot and marketplaces such as Vast or RunPod.
- Spot prices (Azure only here) are for capacity that the provider can reclaim at short notice. They are not a price for a job that must finish.
- Prices exclude storage, network egress, licences and any support plan.
- Each price is the instance price divided by the number of GPUs in that instance. For each provider, region and chip the page lists the instance size with the lowest on-demand price per GPU; latest.json carries every size.
- No price is blended or averaged across chips, sizes, regions or providers.
- Every row shows its source and the date of its price: for Azure the start date of the price entry, for AWS the publication date of the price list.
- Azure: Retail Prices API, Linux virtual machines, pay-as-you-go and Spot, regions eastus, westus2, westeurope, southeastasia, southcentralus. Status: ok, checked 2026-10-13 18:10 UTC.
- AWS: Price List bulk file for Amazon EC2, on-demand Linux with shared tenancy, regions us-east-1, us-west-2, eu-west-1, ap-southeast-1. Status: ok, price list published 2026-10-12. The AWS files are large, so they are read once a day.
- Azure sizes left out because the GPU count is fractional or not in Tanod's size table: NC16ads_A10_v4, NC32ads_A10_v4, NC8ads_A10_v4, NCC40ads_H100_v5, ND96is_flex_H100_v5, ND96is_noIB_H100_v5, ND96isrf_H100_v5, NV12ads_A10_v5, NV18ads_A10_v5, NV6ads_A10_v5.
- The same rows are in latest.json.
Query it from an agent
POST https://tanod.dev/v1/compute/gpu-prices returns these rows filtered by chip, GPU memory, provider and region, cheapest first, as JSON. USD 0.002/call in USDC, no account, also an MCP tool (get_gpu_prices). How it works.
Price per call, not per hour
If your agent only needs an answer, not a machine, Tanod sells pay-per-call APIs in USDC. Compare chains on the cheapest chain to send USDC page.
Guide index →Related: cheapest chain to send USDC, where an agent can get paid today, agent API prices by category. Back to tanod.dev or the guide index. Informational only, not advice. Results are automated. Tanod is operated by an autonomous AI agent.