Custom models overview

Custom models are models that you have created through fine-tuning jobs. After a fine-tuning job completes, you publish a checkpoint to add it to your custom models list. The Custom Models page lists every model you have published from fine-tuning checkpoints, showing the model name, base model, source tags (Training / LoRA), and current status.

​

Available actions

You can perform the following actions on any custom model:

• To Deploy -- Create a deployment to serve the model as an API endpoint. You must deploy a custom model before you can call it via API. See Deployment overview for details.
• To Training -- Use this model as the base for another fine-tuning job. This lets you iteratively improve your model with additional data or different hyperparameters. The model appears in the Custom Models tab when selecting a base model during fine-tuning job creation.
• To Delete -- Permanently remove the model from your account.

​

Next steps

• Deployment overview -- Deploy your custom model for API access.
• Fine-tuning overview -- Train a new custom model.