Isolated deployments
Run your workloads on dedicated serving capacity reserved for your team, separate from the shared, token-based API service.
All inference products
Dedicated Deployments
Your models.Your dedicated capacity.
Run models in an isolated deployment with capacity reserved for your team. Dedicated Deployments bring more control over model weights, configuration and infrastructure for workloads with specific performance or compliance requirements.
Join the waitlist- Isolated deployments
- Custom model weights
- Zero data retention
Built around your requirements.
Shape your deployment around the model, capacity and operating requirements your team needs.
Compliance requirements
Discuss your security, governance and infrastructure requirements with our team to scope a deployment that supports your compliance needs.
Custom weight configurations
Request a deployment for your own model weights or a specific model configuration. Work with our team to assess compatibility and serving requirements.
Zero data retention
Dedicated Deployments operate with zero data retention (ZDR), bringing the same data-retention approach to your isolated environment.
Potential for fixed costs
For steady workloads, dedicated capacity can offer a path to predictable, fixed infrastructure costs. Available terms depend on your deployment and capacity requirements.
Capacity for your workload
Plan the resources your application needs around expected demand and performance targets, with deployment sizing tailored to your use case.
Make room for your workload.
Join the waitlist and tell us about your model, capacity and compliance requirements.
Join the waitlist