Microsoft Azure
Best for Enterprises, OpenAI Integrations, Hybrid Cloud
from$1.00/ hr★ 9.2/10
Replicate
Available NowBest for Serverless Image Generation, LLM API inference, Open-Source Model Hosting
About Replicate
Replicate is an API-first machine learning cloud that lets developers run open-source models with thousands of pre-configured endpoints. When you need to integrate Llama 3 or Flux into an app quickly, Replicate handles the entire underlying GPU infrastructure.
Pros & Cons
Pros
- Vast library of pre-configured open-source models (Llama, Stable Diffusion)
- Simple REST API endpoint for any model
- Automatic scaling including scale-to-zero
- Deploy custom models with Cog easily
Cons
- Cold start times can be high for custom unoptimized models
- Strict API-only interface (no bare server access)
- Pricing per second is slightly higher than raw compute
Ideal Use Cases
AI InferenceFine-TuningImage Generation
| GPU Models | H100, A100 80GB, A100 40GB, A40 |
| GPU Types | A100, H100 |
| Headquarters | San Francisco, CA, USA |
| Founded | 2019 |
| Availability | Available Now |
| Website | replicate.com ↗ |
$0.36/ hour (starting)—$4.14/ hr (max)
💡 Pricing note: Rates shown are indicative. Final pricing depends on GPU model, reservation type (spot vs. on-demand), contract length, and region. Get an exact quote →
US
EU
Starting from
$0.36/hr
Up to $4.14/hr
Frequently Asked Questions
Replicate GPU cloud pricing starts from $0.36/hr depending on GPU type, reservation model (on-demand vs. spot vs. reserved), and region. Use the quote form to get exact pricing for your specific workload.
Replicate offers H100, A100 80GB, A100 40GB, A40 GPU instances. Availability varies by region and configuration. Contact the provider through ComputeStacker for current availability.
Replicate operates data centers in EU West, US East, US West. Choosing a region close to your users minimises latency and can help with data residency compliance requirements.
Use the "Get a Quote" button on this page to submit your GPU requirements. ComputeStacker will forward your request to Replicate and other matching providers. You'll receive proposals within 24 hours — no commitment required.
Replicate offers high-performance GPU infrastructure suitable for large language model training and fine-tuning workloads. For large-scale distributed training, check the Specs tab for NVLink and InfiniBand interconnect availability.
Explore More Options
Similar Providers
Compare Head-to-Head
Alternatives to Replicate
FluidStack
Best for Enterprise AI Training, Multi-Tenant GPU Clusters, Cost-Effective H100 Access
from$0.89/ hr★ 8.7/10
Crusoe Cloud
Best for Environmentally conscious organizations, AI Training
from$1.50/ hr★ 8.9/10