Market report · Live prices
GPU cloud prices
Every GPU machine in every region we track, from our own price index. Each price is the median over the last 24 hours across the region’s zones.
Updated 7 October 2026, 03:54 UTC. AWS: 41 machine types in 23 regions; Google Cloud: 23 machine types in 32 regions.
896 machines priced on spot. Clear filters
| Machine | Region | GPUs | vCPU / RAM | Spot / hr | Spot / GPU-hr | Spot per GB / hr | Spot discount | Spot vs last week | Daily swing |
|---|---|---|---|---|---|---|---|---|---|
| g6e.2xlarge AWS | N. Virginia (us-east-1) | 1× L40S 48 GB | 8 / 64 GB | $2.212 | $2.212 | $0.0461 | 1% | 0% | 0.1% |
| g6e.48xlarge AWS | Seoul (ap-northeast-2) | 8× L40S 48 GB | 192 / 1536 GB | $17.781 | $2.223 | $0.0463 | 52% | −1% | 0.4% |
| g6e.xlarge AWS | Mumbai (ap-south-1) | 1× L40S 48 GB | 4 / 32 GB | $2.235 | $2.235 | $0.0466 | 0% | +38% | 6.5% |
| a2-highgpu-1g Google Cloud | us-west4 | 1× A100 40 GB | 12 / 85 GB | $2.243 | $2.243 | $0.0561 | — | −5% | 2.5% |
| a2-highgpu-2g Google Cloud | us-west4 | 2× A100 40 GB | 24 / 170 GB | $4.485 | $2.243 | $0.0561 | — | −5% | 2.5% |
| a2-highgpu-4g Google Cloud | us-west4 | 4× A100 40 GB | 48 / 340 GB | $8.970 | $2.243 | $0.0561 | — | −5% | 2.5% |
| a2-highgpu-8g Google Cloud | us-west4 | 8× A100 40 GB | 96 / 680 GB | $17.941 | $2.243 | $0.0561 | — | −5% | 2.5% |
| g6e.8xlarge AWS | Ohio (us-east-2) | 1× L40S 48 GB | 32 / 256 GB | $2.249 | $2.249 | $0.0469 | 50% | −1% | 0.3% |
| g4-standard-192 Google Cloud | asia-northeast1 | 4× RTX PRO 6000 96 GB | 192 / 720 GB | $9.013 | $2.253 | $0.0235 | — | 0% | 0.0% |
| g4-standard-384 Google Cloud | asia-northeast1 | 8× RTX PRO 6000 96 GB | 384 / 1440 GB | $18.025 | $2.253 | $0.0235 | — | 0% | 0.0% |
| g4-standard-48 Google Cloud | asia-northeast1 | 1× RTX PRO 6000 96 GB | 48 / 180 GB | $2.253 | $2.253 | $0.0235 | — | 0% | 0.0% |
| g4-standard-96 Google Cloud | asia-northeast1 | 2× RTX PRO 6000 96 GB | 96 / 360 GB | $4.506 | $2.253 | $0.0235 | — | 0% | 0.0% |
| g6e.24xlarge AWS | Ohio (us-east-2) | 4× L40S 48 GB | 96 / 768 GB | $9.035 | $2.259 | $0.0471 | 40% | −1% | 15.6% |
| g5.48xlarge AWS | Montréal (ca-central-1) | 8× A10G 24 GB | 192 / 768 GB | $18.085 | $2.261 | $0.0942 | 0% | 0% | 0.0% |
| a2-megagpu-16g Google Cloud | asia-southeast1 | 16× A100 40 GB | 96 / 1360 GB | $36.196 | $2.262 | $0.0566 | — | 0% | 0.0% |
| p4d.24xlarge AWS | N. Virginia (us-east-1) | 8× A100 40 GB | 96 / 1152 GB | $18.192 | $2.274 | $0.0568 | 17% | +2% | 3.0% |
| g5.16xlarge AWS | Seoul (ap-northeast-2) | 1× A10G 24 GB | 64 / 256 GB | $2.276 | $2.276 | $0.0949 | 55% | −0% | 0.3% |
| g6e.16xlarge AWS | Stockholm (eu-north-1) | 1× L40S 48 GB | 64 / 512 GB | $2.278 | $2.278 | $0.0475 | 72% | +1% | 4.4% |
| g6.8xlarge AWS | ap-south-2 | 1× L4 24 GB | 32 / 128 GB | $2.285 | $2.285 | $0.0952 | 6% | — | 103.5% |
| g4-standard-192 Google Cloud | europe-west4 | 4× RTX PRO 6000 96 GB | 192 / 720 GB | $9.143 | $2.286 | $0.0238 | — | −4% | 1.9% |
| g4-standard-384 Google Cloud | europe-west4 | 8× RTX PRO 6000 96 GB | 384 / 1440 GB | $18.287 | $2.286 | $0.0238 | — | −4% | 1.9% |
| g4-standard-48 Google Cloud | europe-west4 | 1× RTX PRO 6000 96 GB | 48 / 180 GB | $2.286 | $2.286 | $0.0238 | — | −4% | 1.9% |
| g4-standard-96 Google Cloud | europe-west4 | 2× RTX PRO 6000 96 GB | 96 / 360 GB | $4.572 | $2.286 | $0.0238 | — | −4% | 1.9% |
| g6e.16xlarge AWS | N. Virginia (us-east-1) | 1× L40S 48 GB | 64 / 512 GB | $2.290 | $2.290 | $0.0477 | 70% | −1% | 4.3% |
| a2-highgpu-1g Google Cloud | asia-northeast1 | 1× A100 40 GB | 12 / 85 GB | $2.302 | $2.302 | $0.0576 | — | −5% | 2.5% |
| a2-highgpu-1g Google Cloud | asia-northeast3 | 1× A100 40 GB | 12 / 85 GB | $2.302 | $2.302 | $0.0576 | — | −5% | 2.5% |
| a2-highgpu-2g Google Cloud | asia-northeast1 | 2× A100 40 GB | 24 / 170 GB | $4.605 | $2.302 | $0.0576 | — | −5% | 2.5% |
| a2-highgpu-2g Google Cloud | asia-northeast3 | 2× A100 40 GB | 24 / 170 GB | $4.605 | $2.302 | $0.0576 | — | −5% | 2.5% |
| a2-highgpu-4g Google Cloud | asia-northeast1 | 4× A100 40 GB | 48 / 340 GB | $9.210 | $2.302 | $0.0576 | — | −5% | 2.5% |
| a2-highgpu-4g Google Cloud | asia-northeast3 | 4× A100 40 GB | 48 / 340 GB | $9.210 | $2.302 | $0.0576 | — | −5% | 2.5% |
| a2-highgpu-8g Google Cloud | asia-northeast1 | 8× A100 40 GB | 96 / 680 GB | $18.419 | $2.302 | $0.0576 | — | −5% | 2.5% |
| a2-highgpu-8g Google Cloud | asia-northeast3 | 8× A100 40 GB | 96 / 680 GB | $18.419 | $2.302 | $0.0576 | — | −5% | 2.5% |
| g4-standard-192 Google Cloud | europe-west1 | 4× RTX PRO 6000 96 GB | 192 / 720 GB | $9.385 | $2.346 | $0.0244 | — | −4% | 1.8% |
| g4-standard-384 Google Cloud | europe-west1 | 8× RTX PRO 6000 96 GB | 384 / 1440 GB | $18.771 | $2.346 | $0.0244 | — | −4% | 1.8% |
| g4-standard-48 Google Cloud | europe-west1 | 1× RTX PRO 6000 96 GB | 48 / 180 GB | $2.346 | $2.346 | $0.0244 | — | −4% | 1.8% |
| g4-standard-96 Google Cloud | europe-west1 | 2× RTX PRO 6000 96 GB | 96 / 360 GB | $4.693 | $2.346 | $0.0244 | — | −4% | 1.8% |
| g6e.48xlarge AWS | Ohio (us-east-2) | 8× L40S 48 GB | 192 / 1536 GB | $18.777 | $2.347 | $0.0489 | 38% | +1% | 0.4% |
| g4-standard-192 Google Cloud | europe-west10 | 4× RTX PRO 6000 96 GB | 192 / 720 GB | $9.406 | $2.352 | $0.0245 | — | 0% | 0.0% |
| g4-standard-384 Google Cloud | europe-west10 | 8× RTX PRO 6000 96 GB | 384 / 1440 GB | $18.812 | $2.352 | $0.0245 | — | 0% | 0.0% |
| g4-standard-48 Google Cloud | europe-west10 | 1× RTX PRO 6000 96 GB | 48 / 180 GB | $2.352 | $2.352 | $0.0245 | — | 0% | 0.0% |
| g4-standard-96 Google Cloud | europe-west10 | 2× RTX PRO 6000 96 GB | 96 / 360 GB | $4.703 | $2.352 | $0.0245 | — | 0% | 0.0% |
| g4-standard-192 Google Cloud | europe-west2 | 4× RTX PRO 6000 96 GB | 192 / 720 GB | $9.426 | $2.357 | $0.0245 | — | −5% | 2.5% |
| g4-standard-384 Google Cloud | europe-west2 | 8× RTX PRO 6000 96 GB | 384 / 1440 GB | $18.852 | $2.357 | $0.0245 | — | −5% | 2.5% |
| g4-standard-48 Google Cloud | europe-west2 | 1× RTX PRO 6000 96 GB | 48 / 180 GB | $2.357 | $2.357 | $0.0245 | — | −5% | 2.5% |
| g4-standard-96 Google Cloud | europe-west2 | 2× RTX PRO 6000 96 GB | 96 / 360 GB | $4.713 | $2.357 | $0.0245 | — | −5% | 2.5% |
| g4-standard-384 Google Cloud | europe-west8 | 8× RTX PRO 6000 96 GB | 384 / 1440 GB | $18.968 | $2.371 | $0.0247 | — | +10% | 4.7% |
| g4-standard-48 Google Cloud | europe-west8 | 1× RTX PRO 6000 96 GB | 48 / 180 GB | $2.371 | $2.371 | $0.0247 | — | +10% | 4.7% |
| g4-standard-192 Google Cloud | europe-west8 | 4× RTX PRO 6000 96 GB | 192 / 720 GB | $9.484 | $2.371 | $0.0247 | — | +10% | 4.7% |
| g4-standard-96 Google Cloud | europe-west8 | 2× RTX PRO 6000 96 GB | 96 / 360 GB | $4.742 | $2.371 | $0.0247 | — | +10% | 4.7% |
| g6e.4xlarge AWS | Mumbai (ap-south-1) | 1× L40S 48 GB | 16 / 128 GB | $2.378 | $2.378 | $0.0495 | 34% | +6% | 13.9% |
| g6e.4xlarge AWS | N. Virginia (us-east-1) | 1× L40S 48 GB | 16 / 128 GB | $2.385 | $2.385 | $0.0497 | 21% | +3% | 3.5% |
| g5.16xlarge AWS | Frankfurt (eu-central-1) | 1× A10G 24 GB | 64 / 256 GB | $2.386 | $2.386 | $0.0994 | 53% | −2% | 4.1% |
| a2-highgpu-1g Google Cloud | asia-southeast1 | 1× A100 40 GB | 12 / 85 GB | $2.402 | $2.402 | $0.0601 | — | 0% | 0.0% |
| a2-highgpu-2g Google Cloud | asia-southeast1 | 2× A100 40 GB | 24 / 170 GB | $4.804 | $2.402 | $0.0601 | — | 0% | 0.0% |
| a2-highgpu-4g Google Cloud | asia-southeast1 | 4× A100 40 GB | 48 / 340 GB | $9.609 | $2.402 | $0.0601 | — | 0% | 0.0% |
| a2-highgpu-8g Google Cloud | asia-southeast1 | 8× A100 40 GB | 96 / 680 GB | $19.218 | $2.402 | $0.0601 | — | 0% | 0.0% |
| g6e.12xlarge AWS | Stockholm (eu-north-1) | 4× L40S 48 GB | 48 / 384 GB | $9.628 | $2.407 | $0.0502 | 14% | −1% | 45.2% |
| g4dn.16xlarge AWS | Seoul (ap-northeast-2) | 1× T4 16 GB | 64 / 256 GB | $2.410 | $2.410 | $0.1506 | 55% | 0% | 0.0% |
| p4de.24xlarge AWS | Oregon (us-west-2) | 8× A100 80 GB | 96 / 1152 GB | $19.291 | $2.411 | $0.0301 | 30% | −6% | 3.9% |
| g4-standard-192 Google Cloud | asia-southeast1 | 4× RTX PRO 6000 96 GB | 192 / 720 GB | $9.664 | $2.416 | $0.0252 | — | −1% | 0.7% |
| g4-standard-384 Google Cloud | asia-southeast1 | 8× RTX PRO 6000 96 GB | 384 / 1440 GB | $19.327 | $2.416 | $0.0252 | — | −1% | 0.7% |
| g4-standard-48 Google Cloud | asia-southeast1 | 1× RTX PRO 6000 96 GB | 48 / 180 GB | $2.416 | $2.416 | $0.0252 | — | −1% | 0.7% |
| g4-standard-96 Google Cloud | asia-southeast1 | 2× RTX PRO 6000 96 GB | 96 / 360 GB | $4.832 | $2.416 | $0.0252 | — | −1% | 0.7% |
| p5.48xlarge AWS | Stockholm (eu-north-1) | 8× H100 80 GB | 192 / 2048 GB | $19.388 | $2.424 | $0.0303 | 67% | +40% | 8.9% |
| a2-highgpu-1g Google Cloud | me-west1 | 1× A100 40 GB | 12 / 85 GB | $2.424 | $2.424 | $0.0606 | — | 0% | 0.0% |
| a2-highgpu-2g Google Cloud | me-west1 | 2× A100 40 GB | 24 / 170 GB | $4.849 | $2.424 | $0.0606 | — | 0% | 0.0% |
| a2-highgpu-4g Google Cloud | me-west1 | 4× A100 40 GB | 48 / 340 GB | $9.697 | $2.424 | $0.0606 | — | 0% | 0.0% |
| a2-highgpu-8g Google Cloud | me-west1 | 8× A100 40 GB | 96 / 680 GB | $19.395 | $2.424 | $0.0606 | — | 0% | 0.0% |
| g6.16xlarge AWS | Tokyo (ap-northeast-1) | 1× L4 24 GB | 64 / 256 GB | $2.428 | $2.428 | $0.1012 | 51% | +2% | 0.6% |
| g6e.12xlarge AWS | Oregon (us-west-2) | 4× L40S 48 GB | 48 / 384 GB | $9.722 | $2.430 | $0.0506 | 7% | +2% | 4.6% |
| g4-standard-384 Google Cloud | us-west4 | 8× RTX PRO 6000 96 GB | 384 / 1440 GB | $19.497 | $2.437 | $0.0254 | — | −4% | 1.9% |
| g4-standard-48 Google Cloud | us-west4 | 1× RTX PRO 6000 96 GB | 48 / 180 GB | $2.437 | $2.437 | $0.0254 | — | −4% | 1.9% |
| g4-standard-192 Google Cloud | us-west4 | 4× RTX PRO 6000 96 GB | 192 / 720 GB | $9.749 | $2.437 | $0.0254 | — | −4% | 1.9% |
| g4-standard-96 Google Cloud | us-west4 | 2× RTX PRO 6000 96 GB | 96 / 360 GB | $4.874 | $2.437 | $0.0254 | — | −4% | 1.9% |
| g6e.16xlarge AWS | Osaka (ap-northeast-3) | 1× L40S 48 GB | 64 / 512 GB | $2.439 | $2.439 | $0.0508 | 78% | +82% | 27.1% |
| g5.24xlarge AWS | Mumbai (ap-south-1) | 4× A10G 24 GB | 96 / 384 GB | $9.779 | $2.445 | $0.1019 | 0% | 0% | 61.1% |
| g6e.24xlarge AWS | Tokyo (ap-northeast-1) | 4× L40S 48 GB | 96 / 768 GB | $9.856 | $2.464 | $0.0513 | 55% | 0% | 0.2% |
| g4-standard-192 Google Cloud | me-central2 | 4× RTX PRO 6000 96 GB | 192 / 720 GB | $9.863 | $2.466 | $0.0257 | — | 0% | 0.0% |
| g4-standard-384 Google Cloud | me-central2 | 8× RTX PRO 6000 96 GB | 384 / 1440 GB | $19.726 | $2.466 | $0.0257 | — | 0% | 0.0% |
| g4-standard-48 Google Cloud | me-central2 | 1× RTX PRO 6000 96 GB | 48 / 180 GB | $2.466 | $2.466 | $0.0257 | — | 0% | 0.0% |
| g4-standard-96 Google Cloud | me-central2 | 2× RTX PRO 6000 96 GB | 96 / 360 GB | $4.932 | $2.466 | $0.0257 | — | 0% | 0.0% |
| p4de.24xlarge AWS | Tokyo (ap-northeast-1) | 8× A100 80 GB | 96 / 1152 GB | $19.838 | $2.480 | $0.0310 | 47% | −20% | 10.5% |
| p4de.24xlarge AWS | N. Virginia (us-east-1) | 8× A100 80 GB | 96 / 1152 GB | $19.845 | $2.481 | $0.0310 | 28% | −7% | 3.3% |
| a2-ultragpu-1g Google Cloud | us-east4 | 1× A100 80 GB | 12 / 170 GB | $2.502 | $2.502 | $0.0313 | — | 0% | 0.0% |
| a2-ultragpu-2g Google Cloud | us-east4 | 2× A100 80 GB | 24 / 340 GB | $5.004 | $2.502 | $0.0313 | — | 0% | 0.0% |
| a2-ultragpu-4g Google Cloud | us-east4 | 4× A100 80 GB | 48 / 680 GB | $10.008 | $2.502 | $0.0313 | — | 0% | 0.0% |
| a2-ultragpu-8g Google Cloud | us-east4 | 8× A100 80 GB | 96 / 1360 GB | $20.015 | $2.502 | $0.0313 | — | 0% | 0.0% |
| g4-standard-192 Google Cloud | asia-south1 | 4× RTX PRO 6000 96 GB | 192 / 720 GB | $10.031 | $2.508 | $0.0261 | — | +1% | 0.6% |
| g4-standard-384 Google Cloud | asia-south1 | 8× RTX PRO 6000 96 GB | 384 / 1440 GB | $20.063 | $2.508 | $0.0261 | — | +1% | 0.6% |
| g4-standard-48 Google Cloud | asia-south1 | 1× RTX PRO 6000 96 GB | 48 / 180 GB | $2.508 | $2.508 | $0.0261 | — | +1% | 0.6% |
| g4-standard-96 Google Cloud | asia-south1 | 2× RTX PRO 6000 96 GB | 96 / 360 GB | $5.016 | $2.508 | $0.0261 | — | +1% | 0.6% |
| g4-standard-192 Google Cloud | us-east5 | 4× RTX PRO 6000 96 GB | 192 / 720 GB | $10.097 | $2.524 | $0.0263 | — | −4% | 1.9% |
| g4-standard-384 Google Cloud | us-east5 | 8× RTX PRO 6000 96 GB | 384 / 1440 GB | $20.194 | $2.524 | $0.0263 | — | −4% | 1.9% |
| g4-standard-48 Google Cloud | us-east5 | 1× RTX PRO 6000 96 GB | 48 / 180 GB | $2.524 | $2.524 | $0.0263 | — | −4% | 1.9% |
| g4-standard-96 Google Cloud | us-east5 | 2× RTX PRO 6000 96 GB | 96 / 360 GB | $5.049 | $2.524 | $0.0263 | — | −4% | 1.9% |
| g5.24xlarge AWS | Frankfurt (eu-central-1) | 4× A10G 24 GB | 96 / 384 GB | $10.184 | $2.546 | $0.1061 | 0% | +21% | 40.5% |
| g5.48xlarge AWS | Frankfurt (eu-central-1) | 8× A10G 24 GB | 192 / 768 GB | $20.368 | $2.546 | $0.1061 | 0% | 0% | 0.0% |
| g6e.48xlarge AWS | ap-south-2 | 8× L40S 48 GB | 192 / 1536 GB | $20.561 | $2.570 | $0.0535 | 43% | — | 51.5% |
| g5.24xlarge AWS | London (eu-west-2) | 4× A10G 24 GB | 96 / 384 GB | $10.338 | $2.584 | $0.1077 | 0% | 0% | 0.0% |
| g5.48xlarge AWS | London (eu-west-2) | 8× A10G 24 GB | 192 / 768 GB | $20.676 | $2.584 | $0.1077 | 0% | 0% | 0.0% |
Reading the table
- Prices are in US dollars per hour for the whole machine, and per GPU. The machine price includes its CPUs and memory, so a GPU on a large host costs more per GPU-hour.
- Spot per GB divides the spot price by the machine’s total GPU memory: the fair comparison when what you need is room to fit a model.
- Spot vs last week compares the last 24 hours with the same 24 hours a week earlier. Daily swing is how much the daily price varied over the last 7 days.
- A dash means we have no price for it, not that it is free: some clouds publish spot prices only, and some regions sell a machine on one market only.
- A low price is not a promise of capacity. Spot machines can be scarce in exactly the regions where they are cheapest.
Use the data
The whole table is free to use, licensed CC BY 4.0: chart it, build on it and republish it, with credit to Otium.
Get the next report
GPU prices every month, by GPU and by region. One email a month; we’ll ask you to confirm first.
Free. One-click unsubscribe in every issue.
Let the price come to you.
Otium is in closed alpha. Set a deadline for your batch AI work, and it waits for the lowest-cost compute available before then, so you don’t have to watch these prices yourself. Leave your email and we’ll send an invite when a spot opens.
No credit card, no spam — one email when your invite is ready.
Closed alpha — onboarding is gated while we calibrate. Already invited? Sign in.