Paperspace GPU pricing 2026
Notebook-friendly GPU cloud (Gradient for ML, Core for general VMs). Acquired by DigitalOcean in 2023; offering increasingly aligned with DigitalOcean GPU Droplets.
Published rate card
Per-GPU per-hour rates pulled from the vendor pricing page linked above, Last verified June 2026. Rates exclude storage, egress, and any managed-service uplift. Reserved-capacity contracts typically improve on these rates.
| GPU | Configuration | Per GPU per hour |
|---|---|---|
| H100 80GB | Per-GPU on 8x | $5.950 |
| A100 80GB | Per-GPU | $3.180 |
| A100 40GB | Per-GPU | $1.830 |
| V100 32GB | Per-GPU | $1.620 |
| RTX A6000 | Per-GPU | $1.890 |
| RTX 4000 | Per-GPU | $0.560 |
Hidden costs to watch
- Now operated by DigitalOcean (acquired 2023); pricing converging with DigitalOcean GPU Droplets.
- Storage on Gradient and Core billed by GB-month and is a meaningful share of low-end workloads.
- H100 capacity is constrained; A100 and L40 are the practical training tiers.
What Paperspace is best for
Solo ML practitioners, education, and notebook-style workflows.
See the GPU cloud buying guideWorked example
Acme Vision Co. (illustrative example, not a real company) needs to train a 10 billion-parameter vision-language model on a fixed 8-GPU cluster for 30 days at 18 hours per day duty cycle.
At Paperspace's cheapest published rate of $0.56 per GPU-hour, the run costs $2,419 for raw GPU compute, before storage, egress, and MLOps overhead. Add the typical 25 percent year-one uplift and the modelled spend is $3,024. Use the calculator on the homepage to model your own GPU class, cluster size, and duty cycle.
Last verified June 2026. Paperspace rates change frequently. Always obtain a vendor quote before purchase.
Visit Paperspace pricing page