Paperspace offers an array of virtual machines for various use cases. If you're looking for Gradient Instances, click here.
Standard GPU (Windows Only)
Use case: Optimized for general purpose desktop computing. This family is great for a range of use-cases like Office, software development, and 3D CAD.
GPU type: Standard
Features:
| Model | vCPU | RAM | GPU |
| Air | 2 | 4 GB | 512MB |
| Standard | 4 | 8 GB | 512MB |
| Advanced | 6 | 16 GB | 2GB |
| Pro | 8 | 32 GB | 4GB |
- Intel® Xeon® E5-2630 v3 with base processor frequency of 2.4GHz and max turbo frequency of 3.20 GHz
- NVIDIA GRID K1 GPU
Dedicated GPU
GPU+ (Windows & Linux)
Use case: Optimized for graphics-intensive applications, machine learning and general-purpose GPU compute.
GPU type: Dedicated
Features:
| Model | vCPU | RAM | GPU |
| GPU+ | 8 | 30 GB | 8 GB |
- Intel® Xeon® E5-2623 v4 with base processor frequency of 2.60 GHz and max turbo frequency of 3.20 GHz
US Regions
- NVIDIA M4000
- Architecture: Maxwell 2
- CUDA Cores: 1664
- GPU Memory: 8 GB GDDR5
- TeraFLOPS: 2.6
US & Europe Regions
- NVIDIA P4000
- Architecture: Pascal
- CUDA Cores: 1792
- GPU Memory: 8 GB GDDR5
- TeraFLOPS: 5.3
P5000 (Windows & Linux)
Use case: Optimized for heavy graphics processing, machine learning, and general-purpose GPU compute.
GPU type: Dedicated
Features:
| Model | vCPU | RAM | GPU |
| P5000 | 8 | 30 GB | 16 GB |
- Intel® Xeon® E5-2623 v4 with base processor frequency of 2.60 GHz and max turbo frequency of 3.20 GHz
- NVIDIA P5000
- Architecture: Pascal
- CUDA Cores: 2560
- GPU Memory: 16 GB GDDR5X
- TeraFLOPS: 9
P6000 (Windows & Linux)
Use case: Optimized for extreme graphics processing, machine learning, and high-end GPU compute.
GPU type: Dedicated
Features:
| Model | vCPU | RAM | GPU |
| P6000 | 8 | 30 GB | 24 GB |
- Intel® Xeon® E5-2623 v4 with base processor frequency of 2.60 GHz and max turbo frequency of 3.20 GHz
- NVIDIA P6000
- Architecture: Pascal
- CUDA Cores: 3840
- GPU Memory: 24 GB GDDR5X
- TeraFLOPS: 12
V100 (Linux Only)
Use case: The most powerful card offered today (latest Volta generation), optimized specifically for Deep Learning and HPC.
GPU type: Dedicated
Features:
| Model | vCPU | RAM | GPU |
| V100 | 8 | 30 GB | 16 GB HBM2 |
- Intel® Xeon® E5-2623 v4 with base processor frequency of 2.60 GHz and max turbo frequency of 3.20 GHz
- NVIDIA V100
- Architecture: Volta
- CUDA Cores: 5120
- GPU Memory: 16 GB HBM2
- TeraFLOPS: 14
- Memory Bandwidth: 900 GB/s
- Tensor Performance: 112 TFLOPS
C Class (Linux Terminal Only)
Use case: Optimized for web servers. This family provides a balance of compute, memory, and network performance and it is a good choice for many applications.
GPU type: None
Options:
| Model | vCPU | RAM |
| C1 | 1 | 512 MB |
| C2 | 1 | 1 GB |
| C3 | 2 | 2 GB |
| C4 | 2 | 4 GB |
| C5 | 4 | 8 GB |
| C6 | 8 | 16 GB |
| C7 | 12 | 30 GB |
| C8 | 16 | 60 GB |
| C9 | 24 | 120 GB |
| C10 | 32 | 244 GB |
- Intel® Xeon® E5-2630 v3 with base processor frequency of 2.4GHz and max turbo frequency of 3.20 GHz
Features
Storage
All machine types
SSD drives are used for the root volume of the machine.
Network
All machine types
Speeds for all instance types are 1 Gigabit up/down