Aceus GPU Pricing & Specifications
At Aceus you can choose a wide variety of GPU powered instances for your AI model training and deployments. We offer 2 kinds of instances
- On demand instances, you can launch them in seconds.
- Bare metal server for 30 or more days.
- Free bandwidth
Bare metal server
For reserving the below servers, please reach out to us using the chat option or drop us an email at hello@Aceus.com.
GPU Type | vCPUS | Ram | Storage | 7 Days | 30 Days | 90 Days | 183 Days | 365 Days | NVlink |
---|---|---|---|---|---|---|---|---|---|
1x A100 80 GB/SXM | 16 | 115GB | 1500GB | $380 | $1440 | $4000 | $7600 | $13500 | Yes |
2x A100 80 GB/SXM | 32 | 230GB | 3000GB | $760 | $2880 | $8000 | $15200 | $27000 | Yes |
4x A100 80 GB/SXM | 64 | 460GB | 6000GB | $1520 | $5760 | $16000 | $30400 | $54000 | Yes |
2x A30 48 GB | 32 | 180GB | 1280GB | - | $1110 | $3250 | $6340 | $12000 | No |
4x A30 96 GB | 64 | 360GB | 2560GB | - | $2220 | $6500 | $12680 | $24000 | No |
On Demand instances
On demand instances comes preconfigured with
- Jupyter Lab
- VS Code
- SSH access
- API endpoint (Useful for deploying apps like Gradio, FastAPI servers)
- Free bandwidth
- Pay per minute
- Choose upto 8 GPUs on a single instance
Nvidia Ada Lovelace GPU
The RTX6000 Ada GPU delivers superior performance compared to the previous-generation Ampere GPUs, outperforming predecessors such as the A6000 and A5000.
GPU | vCPUS | Ram | Spot | Reserved | Weekly | Monthly |
---|---|---|---|---|---|---|
RTX6000 Ada - 48GB | 32 | 128GB | $0.79/hr | $1.14/hr | $180/week | $680/month |
Nvidia Ampere GPU
Choose from A100, A6000 and A5000 GPU, provides up to 20X higher performance over the prior generation. Ampere GPUs are also faster than V100 and Quadro based GPUs.
GPU | vCPUS | Ram | Spot | Reserved | Weekly | Monthly |
---|---|---|---|---|---|---|
A100 - 40GB | 7 | 32GB | $0.79/hr | $1.49/hr | $230/week | $860/month |
A6000 - 48GB | 7 | 32GB | $0.59/hr | $0.99/hr | $155/week | $640/month |
A5000 - 24GB | 7 | 32GB | $0.39/hr | $0.59/hr | $95/week | $400/month |
Nvidia Quadro GPU
Choose from RTX 6000 and RTX 5000, provides similar performance to an V100 GPU.
GPU | vCPUS | Ram | Spot | Reserved | Weekly | Monthly |
---|---|---|---|---|---|---|
RTX 5000 - 16GB | 7 | 32GB | $0.19/hr | $0.49/hr | $75/week | $315/month |
Spot instances
Spot instances or preemptible instances lets you use spare GPUs at a discounted price and save upto 60% cost. However these instances can be auto-paused when the demand increases.
All the data stored on the disk will remain intact.
These instances are well suited for
- Learning.
- When you have implemented checkpoints, and you want to save on costs.
- Scaling API workloads that are not mission critical.
On Demand Weekly/Monthly duration
Get upto 5% discounts for weekly usage and upto 10% discounts on monthly usage. Instances that are launched for a week/month cannot be paused or destroyed during that duration. Post the duration, instances will continue at the discounted prices till you choose to pause/destroy.
Storage
Storage costs $0.1/GB per month and you can choose upto 500GB per instance.
If you need more storage, please reach out to us using the chat option or drop us an email at hello@Aceus.com.