Skip to Content

Model Management

Browse, deploy, train, evaluate, and monitor AI models.

Model Catalog

/dashboard/models — browse available models with provider, context window, and pricing. Filter by provider or capability.

Deployments

/dashboard/models/deployments — manage active model deployments:

  • Deploy new models with GPU type and region config
  • Monitor deployment status (running, deploying, failed)
  • Metric cards: total deployments, running, failed, estimated cost
  • Restart or delete deployments

Training

/dashboard/models/training — manage training jobs:

  • Start training with algorithm selection (APO, GRPO, closed-loop, custom)
  • Configure hyperparameters, datasets, and training budget
  • Monitor job status and progress
  • Export training data in multiple formats

Evaluation

/dashboard/models/evaluation — evaluate model performance:

  • Create evaluation suites with test cases
  • Run evaluations with configurable metrics
  • View results with pass/fail breakdowns
  • Compare evaluation runs over time

Prediction

/dashboard/models/prediction — test model predictions:

  • Submit prediction requests with custom inputs
  • View response quality and latency
  • Train-on-predict for active learning workflows

Analytics

/dashboard/models/analytics — model performance dashboards:

  • Request volume, latency percentiles, error rates
  • Cost tracking per model
  • Usage trends over time
Home