Skip to main content
7d ago
GPU Tracker — Compare Cloud GPU PricesGPU Tracker

Replicate GPU Pricing

4 GPU instances across 1 regions.4 GPU models available — from $0.81/hr.

Visit Replicate
ServerlessInferencePer-second billing

Model hosting platform that lets you run open-source ML models in the cloud with a single API call. Replicate handles infrastructure while you pay per second of GPU time.

Strengths
  • Run any open-source model with one API call
  • Per-second billing — very cost-effective for bursty workloads
  • Community model library with thousands of models
  • Custom model deployment via Cog
Considerations
  • Not suitable for training — inference only
  • No persistent storage
  • Cold start latency for infrequently-used models
Stop to pause billing
Free egress
Quick Answer•Updated Apr 12, 7:00 AM•Methodology

Replicate currently lists 4 GPU instances across 4 GPU models and 1 regions. Pricing starts at $0.81/hr, while the median listing price is $4.14/hr. Compare by model, commitment type, and region before treating the cheapest row as the best choice.

Starting at
$0.81/hr
Median
$4.14/hr
Models
4
Spot share
0%
T4 pricingA40 pricingA100 40GB pricingA100 80GB pricingMarket trends
Starting at
$0.81/hr
cheapest instance
GPU Models
4
available
Instances
4
total
Regions
1
covered

GPU Models at Replicate

T4$0.810/hrA40$2.61/hrA100 40GB$4.14/hrA100 80GB$5.04/hr

All Replicate GPU Instances

4 results
GPU ModelInstanceCountVRAMRegionTypePrice/hr$/GPU/hr
T4T4-replicate1×16GBUS-WestServerless$0.8100—Rent
A40A40-replicate1×48GBUS-WestServerless$2.6100—Rent
A100 40GBA100-40GB-replicate1×40GBUS-WestServerless$4.1400—Rent
A100 80GBA100-80GB-replicate1×80GBUS-WestServerless$5.0400—Rent

Compare Other Providers

Vast.aiVerdaVultrAzureRunPodGCPAWSE2E NetworksShadeformHyperstackSalad CloudGenesis CloudDatacrunchFluidStackBeamTensorDockCloudRiftCrusoeLightning.aiNovitaJarvis LabsThunder ComputeLeaderGPULinodeCerebriumMassed ComputePaperspaceSesterceModalOCINebiusScalewayContaboOVHcloudCherry ServersNscaleLambda LabsGreen AI CloudHetznerCivoKoyebCirrascaleFly.ioCudo ComputeLatitude.shTogether AIOblivusVoltage ParkTensorWaveGcoreFal.aiDigitalOceanCoreWeave

Replicate vs … (Head-to-Head)

Replicate vs Vast.aiReplicate vs VerdaReplicate vs VultrReplicate vs AzureReplicate vs RunPodReplicate vs GCPReplicate vs AWSReplicate vs E2E NetworksReplicate vs ShadeformReplicate vs HyperstackReplicate vs Salad CloudReplicate vs Genesis CloudReplicate vs DatacrunchReplicate vs FluidStackReplicate vs BeamReplicate vs TensorDockReplicate vs CloudRiftReplicate vs CrusoeReplicate vs Lightning.aiReplicate vs NovitaReplicate vs Jarvis LabsReplicate vs Thunder ComputeReplicate vs LeaderGPUReplicate vs Linode

Replicate GPU Cloud — FAQ

How much does Replicate charge for GPUs?

Replicate GPU instances start from $0.81/hr. The average price is $3.15/hr. Prices depend on GPU model, region, and commitment type (on-demand vs spot).

What GPU models does Replicate offer?

Replicate offers 4 GPU models: T4, A40, A100 40GB, A100 80GB. Browse the full list above to compare prices per model.

Where can I see billing assumptions and risk methodology?

GPU Tracker’s pricing comparisons are paired with true cost and risk signals. Read the methodology page for how refresh cadence, cost assumptions, and reliability indicators are defined.