Paperspace alternatives
Every GPU compute provider this site tracks, ordered by iScore — the same peers Paperspace's own score is measured against. The line under each name is where its sourced facts differ from Paperspace's; where a full comparison exists, the row links to it.
- RunPodEntry paid plan: No plan tiers: a new account can start with as little as $10 in prepaid credits, and depl…
- DeepInfraFree tier: No vs Yes · Kind of free offer: No free offer vs Free tier
- Novita AIEntry paid plan: No plan fee — credit is prepaid, and a manual top-up must be greater than $10.
- Nebius AI CloudEntry paid plan: No plan tiers: usage is paid as you go in USD, and the first payment has a $25 minimum.
- TensorDockEntry paid plan: No plan tiers: deposit funds from as little as $5, drawn down as servers run.
- LambdaEntry paid plan: Pay-as-you-go per GPU-hour rather than a monthly plan: the cheapest self-serve instance i…
- BasetenKind of free offer: Free credit vs Free tier
- CoreWeavePricing model: Four CKS capacity plans - Flex Reservations, Reserved Instances, Spot Instances and On-De…
- Crusoe CloudEntry paid plan: No plan tiers or setup fees: pay-as-you-go by the hour. The cheapest GPU is the L40S 48GB…
- Vast.aiFree tier: No vs Yes · Kind of free offer: No free offer vs Free tier
- HyperstackKind of free offer: No free offer vs Free tier
- Fireworks AIFree tier: No vs Yes · Kind of free offer: Free credit vs Free tier
Where Paperspace stands against these 12
Paperspace's own sourced figures, placed among the alternatives above that document the same thing in the same unit. A rank here is one criterion, not a verdict; the iScore above weighs all of them.
- Free tier
- Yes. 1 of the 7 alternatives that document it has one: Novita AI.
- Egress over allowance
- FreeJoint cheapest of the 7 that document it. Level with: CoreWeave — Free.
- Regions
- Three datacenters: NY2 (near New York City), CA1 (near Santa Clara, CA) and AMS1 (near Amsterdam)Joint 4th highest of the 6 that document it. Highest: RunPod — 31 regions across the US, Europe, Asia and Australia for on-demand GPU instances.
- Uptime SLA
- 99.99% uptime SLA stated for network, power, virtual server and storage availabilityHighest of the 3 that document it. Next: Crusoe Cloud — 99.5% monthly uptime per GPU-enabled VM, with service credits of 10%, 25% or 100% of that VM's monthly bill…
- H100 SXM, per GPU-hour
- $5.95 per hour for a single H100 80GB machine (20 vCPU, 250 GB RAM); H100x8 is $47.60 an hour.…9th cheapest of the 12 that document it. Cheapest: DeepInfra — $2.20 per GPU-hour for a dedicated H100 80GB — below every dedicated cloud in this dataset.
- Cheapest GPU, per hour
- $0.45 per hour for GPU+ (M4000) with 8GB; P4000 is $0.51 and RTX4000 $0.564th cheapest of the 8 that document it. Cheapest: Vast.ai — $0.02 per hour for a GTX 1660 S with 6GB, the floor from the cheapest host; its median is $0.05.
- Largest GPU offered
- 80GB per GPU on H100 (HBM3) and A100-80G (HBM2); H100x8 carries 640GB across eight GPUsJoint 11th highest of the 12 that document it. Highest: Crusoe Cloud — 288GB per GPU on AMD MI355X, which is Contact sales with no on-demand price.
A rank whose 80% interval overlaps its neighbour's is an ordering, not a finding; the intervals are on each provider's page and in /scores.json.