Skip to main content
NexusAPI is usage-based: text models can charge for input, output, and cached tokens; image, audio, video, or task models can charge per image, duration, task, or another upstream unit. The live details in the Model Marketplace take precedence over this guide.

​
Model Marketplace

Use the Model Marketplace in the console to see current models, prices, and token-group availability:
  1. Sign in to the NexusAPI console, then open the Model Marketplace.
  2. Filter by model provider and available token group.
  3. Review the model ID, price, description, and supported endpoint types for that group.
NexusAPI Model Marketplace, filterable by provider and token group

​
Choose the right token group

The token group selected when you create a key determines which models and prices the key can use.
  • A model can appear in more than one group; choose by the target model, not only by the group name.
  • Filter the Model Marketplace by the group and confirm the exact model ID.
  • If one project needs different classes of models, separate API keys make limits and reconciliation easier.
Selecting a token group when creating a NexusAPI token

​
What makes up a charge

  • Text models: input tokens, output tokens, and cache-related usage can be priced separately.
  • Image models: size, quality, quantity, and model-specific options can affect the charge.
  • Audio, video, and task models: a charge can depend on duration, task type, or an upstream billing unit.
Use the individual charge in the console usage log and the live Model Marketplace price to reconcile a request. Do not apply token-price rules to image or task models.

​
Check usage

  • View each request in the console usage log.
  • Read usage and billing for field meanings, cache usage, and error records.
  • For programmatic reconciliation, use the usage-log API. It uses a system access token, not a model API key.