Model Lifecycle

Every model on the Nebul Inference API sits in one of three lifecycle stages. Knowing which stage a model is in tells you how stable it is, what kind of support to expect, and whether it's ready for your production workloads.

Lifecycle Stages

Production

Production models are ready for real workloads and run on dedicated GPU infrastructure. They've been through performance validation, so you can build on them with confidence.

• Stable API surface and model behavior
• Full monitoring and alerting
• No breaking changes without advance notice
• Recommended for production use

Preview

Preview models are open to all customers while validation is still underway. They give you early access to new capabilities, with a few trade-offs while we finish testing.

• Model weights or configurations may be updated
• Brief unavailability is possible during infrastructure maintenance
• Will be promoted to Production or deprecated once validation is complete

Early Access

Early Access models are shared with select customers on request, ahead of general availability. They typically run on experimental infrastructure while we actively tune them, and we make them available so you can evaluate and give feedback early.

• Access is granted per organization
• Runs on experimental or non-production infrastructure
• May be interrupted, restarted, or reconfigured without notice
• Model weights, quantization, and serving parameters can change frequently
• Not recommended for production use
• May be deprecated without a migration path

Important

Early Access models are meant for evaluation and experimentation only. Do not use them for production workloads.

How to Tell Which Stage a Model Is In

Check the Model Catalog: models in preview carry a Preview tag, and Production models carry no tag at all. Early Access models are called out in the model description, and we'll let you know directly when access is first granted.

Requesting Access

Want Early Access to a specific model, or think a Preview model is ready for Production? Reach out through the Support Portal or contact your account team. We're happy to talk it through.

Deprecation Policy

When we schedule a model for deprecation, here's what to expect:

1. Production models: at least 14 days' notice before removal
2. Preview models: we aim to give the same 14 days' notice, wherever possible
3. Early Access models: may be removed at any time without notice

We typically reroute a deprecated model to a similar one at a comparable price. Deprecated models stay callable until their removal date, so you have time to migrate. We recommend switching to the successor model as soon as deprecation is announced.