Title here
Summary here
GPU
Simple and Affordable Pricing
All instances are Dedicated GPUs and backed with NVMe storage.
Pricing Table
Caution
To gain access to our GPUs, please reach out to us at [email protected] to request an increase in your quota.
Did you know?
Each GPU Instance has 55GB of disk per vCPU
eg: 16 * 55GiB = 880 GiB
L40s class
These instances are powered by Nvidia L40s. These instances are good for running LLMs using GPU inference like llama3.2-vision:11b-instruct-fp16, nemotron:70b-instruct-q4_K_M, qwen2.5-coder:32b-base-q4_K_M, etc.
| Instance | vCPU | Memory | vRAM | Reserved | On-Demand |
|---|---|---|---|---|---|
| l40s.1x | 16 vCPU | 128 GiB | 48 GiB | $0.80/hr | $1.55/hr |
| l40s.2x | 32 vCPU | 256 GiB | 96 GiB | $1.60/hr | $3.10/hr |
| l40s.4x | 64 vCPU | 512 GiB | 192 GiB | $3.20/hr | $6.20/hr |
MI300X class
These instances are powered by AMD MI300X. These instances are good for running LLMs using Multi GPU inference like llama3.2-vision:11b-instruct-fp16, nemotron:70b-instruct-q4_K_M, qwen2.5-coder:32b-base-q4_K_M, etc.
| Instance | vCPU | Memory | vRAM | Reserved | On-Demand |
|---|---|---|---|---|---|
| mi300x.1x | 16 vCPU | 128 GiB | 192 GiB | $2.90/hr | $5.62/hr |
| mi300x.8x | 128 vCPU | 1024 GiB | 1536 GiB | $23.00/hr | $29.90/hr |
H100 class
These instances are powered by Nvidia H100 SXM. These instances are good for running LLMs using Multi GPU inference like llama3.2-vision:11b-instruct-fp16, nemotron:70b-instruct-q4_K_M, qwen2.5-coder:32b-base-q4_K_M, etc.
| Instance | vCPU | Memory | vRAM | Reserved | On-Demand |
|---|---|---|---|---|---|
| h100.1x | 16 vCPU | 128 GiB | 80 GiB | $2.00/hr | $3.88/hr |
| h100.8x | 128 vCPU | 1024 GiB | 640 GiB | $16/hr | $20.8/hr |