Premium GPU Instances
NVIDIA B300 virtual machines
Built on NVIDIA Blackwell Ultra with 5th Gen AMD EPYC Turin processors and NVLink v5.
NVIDIA’s latest hardware, designed to further accelerate LLM and MoE inference compared to its predecessor.
| GPU model | Instance name | CPU | RAM | VRAM | Pay As You Go Price |
|---|---|---|---|---|---|
| 8x B300 SXM6 | 8B300.240V | 240 | 2200 | 2100 | $39.60/h |
| 4x B300 SXM6 | 4B300.120V | 120 | 1100 | 1050 | $19.80/h |
| 2x B300 SXM6 | 2B300.60V | 60 | 550 | 525 | $9.90/h |
| 1x B300 SXM6 | 1B300.30V | 30 | 275 | 262 | $4.95/h |
- Pricing per GPU
- $4.95/h Pay As You Go
- $1.24/h Spot
- $4.85/h -2%1 month
- $4.55/h -8%1 year
- $3.71/h -25%2 years
NVIDIA B200 virtual machines
Built on NVIDIA Blackwell with NVLink v5.
Promised to deliver breakthrough performance for large-scale AI, powering the most demanding training and inference workloads.
| GPU model | Instance name | CPU | RAM | VRAM | Pay As You Go Price |
|---|---|---|---|---|---|
| 8x B200 SXM6 | 8B200.240V | 240 | 1472 | 1440 | $30.32/h |
| 4x B200 SXM6 | 4B200.120V | 120 | 736 | 720 | $15.16/h |
| 2x B200 SXM6 | 2B200.60V | 60 | 368 | 360 | $7.58/h |
| 1x B200 SXM6 | 1B200.30V | 30 | 184 | 180 | $3.79/h |
- Pricing per GPU
- $3.79/h Pay As You Go
- $0.95/h Spot
- $3.71/h -2%1 month
- $3.49/h -8%1 year
- $2.84/h -25%2 years
NVIDIA H200 virtual machines
Built on NVIDIA Hopper with 4th Gen AMD Genoa processors and NVLink v4.
The 8H200 VM matches the vCPU count of the 4H200 while delivering higher performance by using dedicated physical cores instead of hyper-threads.
| GPU model | Instance name | CPU | RAM | VRAM | Pay As You Go Price |
|---|---|---|---|---|---|
| 8x H200 SXM5 | 8H200.141S.176V | 176 | 1450 | 1128 | $23.92/h |
| 4x H200 SXM5 | 4H200.141S.176V | 176 | 740 | 564 | $11.96/h |
| 2x H200 SXM5 | 2H200.141S.88V | 88 | 370 | 282 | $5.98/h |
| 1x H200 SXM5 | 1H200.141S.44V | 44 | 182 | 141 | $2.99/h |
- Pricing per GPU
- $2.99/h Pay As You Go
- $1.50/h Spot
- $2.93/h -2%1 month
- $2.75/h -8%1 year
- $2.24/h -25%2 years
NVIDIA H100 virtual machines
Built on NVIDIA Hopper with 4th Gen AMD Genoa processors and NVLink v4.
Designed to deliver exceptional performance for training and inference at scale.
| GPU model | Instance name | CPU | RAM | VRAM | Pay As You Go Price |
|---|---|---|---|---|---|
| 8x H100 SXM5 | 8H100.80S.176V | 176 | 1480 | 640 | $18.32/h |
| 4x H100 SXM5 | 4H100.80S.176V | 176 | 740 | 320 | $9.16/h |
| 2x H100 SXM5 | 2H100.80S.80V | 80 | 370 | 160 | $4.58/h |
| 1x H100 SXM5 | 1H100.80S.32V | 32 | 185 | 80 | $2.29/h |
| 1x H100 SXM5 | 1H100.80S.30V | 30 | 120 | 80 | $2.29/h |
- Pricing per GPU
- $2.29/h Pay As You Go
- $0.99/h Spot
- $2.24/h -2%1 month
- $2.11/h -8%1 year
- $1.72/h -25%2 years
NVIDIA A100 virtual machines
Powered by 2nd Gen AMD EPYC Rome processors and NVLink v3, suitable for demanding AI and HPC applications.
| GPU model | Instance name | CPU | RAM | VRAM | Pay As You Go Price |
|---|---|---|---|---|---|
| 8x A100 SXM4 | 8A100.176V | 176 | 960 | 640 | $10.32/h |
| 4x A100 SXM4 | 4A100.88V | 88 | 480 | 320 | $5.16/h |
| 2x A100 SXM4 | 2A100.44V | 44 | 240 | 160 | $2.58/h |
| 1x A100 SXM4 | 1A100.22V | 22 | 120 | 80 | $1.29/h |
| 8x A100 SXM4 | 8A100.40S.176V | 176 | 960 | 320 | $5.77/h |
| 1x A100 SXM4 | 1A100.40S.22V | 22 | 120 | 40 | $0.72/h |
- Pricing per GPU
- $1.29/h Pay As You Go
- $0.58/h Spot
- $1.26/h -2%1 month
- $1.19/h -8%1 year
- $0.97/h -25%2 years
NVIDIA RTX PRO 6000 virtual machines
Powered by 5th Gen AMD EPYC Turin processors, designed to accelerate AI Inference, 3D rendering and real-time virtualization with exceptional efficiency.
| GPU model | Instance name | CPU | RAM | VRAM | Pay As You Go Price |
|---|---|---|---|---|---|
| 8x RTX PRO 6000 | 8RTXPRO6000.240V | 240 | 720 | 768 | $11.12/h |
| 4x RTX PRO 6000 | 4RTXPRO6000.120V | 120 | 360 | 384 | $5.56/h |
| 2x RTX PRO 6000 | 2RTXPRO6000.60V | 60 | 180 | 192 | $2.78/h |
| 1x RTX PRO 6000 | 1RTXPRO6000.30V | 30 | 90 | 96 | $1.39/h |
- Pricing per GPU
- $1.39/h Pay As You Go
- $0.35/h Spot
- $1.36/h -2%1 month
- $1.28/h -8%1 year
- $1.04/h -25%2 years
NVIDIA L40S virtual machines
Powered by 3rd Gen AMD EPYC Milan processors, optimized for training ML models, single-GPU training, and other 31-16-8 bit operations.
| GPU model | Instance name | CPU | RAM | VRAM | Pay As You Go Price |
|---|---|---|---|---|---|
| 8x L40S | 8L40S.160V | 160 | 480 | 384 | $7.31/h |
| 4x L40S | 4L40S.80V | 80 | 240 | 192 | $3.66/h |
| 2x L40S | 2L40S.40V | 40 | 120 | 96 | $1.83/h |
| 1x L40S | 1L40S.20V | 20 | 60 | 48 | $0.91/h |
- Pricing per GPU
- $0.91/h Pay As You Go
- $0.27/h Spot
- $0.90/h -2%1 month
- $0.84/h -8%1 year
- $0.69/h -25%2 years
NVIDIA RTX 6000 ADA virtual machines
Powered by 3rd Gen AMD EPYC Milan processors, suitable for working with the largest 3D models, render images, simulation and AI datasets.
| GPU model | Instance name | CPU | RAM | VRAM | Pay As You Go Price |
|---|---|---|---|---|---|
| 8x RTX 6000 Ada | 8RTX6000ADA.80V | 80 | 480 | 384 | $6.61/h |
| 4x RTX 6000 Ada | 4RTX6000ADA.40V | 40 | 240 | 192 | $3.31/h |
| 2x RTX 6000 Ada | 2RTX6000ADA.20V | 20 | 120 | 96 | $1.65/h |
| 1x RTX 6000 Ada | 1RTX6000ADA.10V | 10 | 60 | 48 | $0.83/h |
- Pricing per GPU
- $0.83/h Pay As You Go
- $0.25/h Spot
- $0.81/h -2%1 month
- $0.76/h -8%1 year
- $0.62/h -25%2 years
NVIDIA RTX A6000 virtual machines
Powered by 2nd Gen AMD EPYC Rome processors, ideal for AI Inference, 3D rendering, and professional visualization workloads.
Equipped with PCle Gen4 for enhanced GPU communication.
| GPU model | Instance name | CPU | RAM | VRAM | Pay As You Go Price |
|---|---|---|---|---|---|
| 8x RTX A6000 | 8A6000.80V | 80 | 480 | 384 | $3.92/h |
| 4x RTX A6000 | 4A6000.40V | 40 | 240 | 192 | $1.96/h |
| 2x RTX A6000 | 2A6000.20V | 20 | 120 | 96 | $0.98/h |
| 1x RTX A6000 | 1A6000.10V | 10 | 60 | 48 | $0.49/h |
- Pricing per GPU
- $0.49/h Pay As You Go
- $0.15/h Spot
- $0.48/h -2%1 month
- $0.45/h -8%1 year
- $0.37/h -25%2 years
NVIDIA V100 virtual machines
Powered by XEON Scalable 4214R processors and NVLink v2, ideal for AI training, Inference, and high-performance computing workloads.
| GPU model | Instance name | CPU | RAM | VRAM | Pay As You Go Price |
|---|---|---|---|---|---|
| 8x Tesla V100 | 8V100.48V | 48 | 180 | 128 | $1.10/h |
| 4x Tesla V100 | 4V100.20V | 20 | 90 | 64 | $0.55/h |
| 2x Tesla V100 | 2V100.10V | 10 | 45 | 32 | $0.28/h |
| 1x Tesla V100 | 1V100.6V | 6 | 23 | 16 | $0.14/h |
- Pricing per GPU
- $0.14/h Pay As You Go
- $0.07/h Spot
- $0.14/h -2%1 month
- $0.13/h -8%1 year
- $0.10/h -25%2 years
NVIDIA CPU virtual machines
Powered by AMD EPYC Rome, Milan, or Genoa, delivering exceptional core density, memory bandwidth, and reliability for high-performance cloud workloads.
| CPU model | Instance name | CPU | RAM | Pay As You Go Price |
|---|---|---|---|---|
| AMD Genoa | CPU.360V.1440G | 360 | 1440 | $2.51/h |
| AMD Genoa | CPU.180V.720G | 180 | 720 | $1.26/h |
| AMD Rome/Milan | CPU.120V.480G | 120 | 480 | $0.84/h |
| AMD Rome/Milan | CPU.96V.384G | 96 | 384 | $0.67/h |
| AMD Rome/Milan | CPU.64V.256G | 64 | 256 | $0.45/h |
| AMD Rome/Milan | CPU.32V.128G | 32 | 128 | $0.22/h |
| AMD Rome/Milan | CPU.16V.64G | 16 | 64 | $0.11/h |
| AMD Rome/Milan | CPU.8V.32G | 8 | 32 | $0.06/h |
| AMD Rome/Milan | CPU.4V.16G | 4 | 16 | $0.03/h |
- Pricing per CPU unit
- $0.03/h Pay As You Go
- $0.01/h Spot
- $0.03/h -2%1 month
- $0.03/h -8%1 year
- $0.02/h -25%2 years