Skip to main content

Model API overview

Choose a model, make a request, and inspect the response and usage.

Start here

Use the Acosmi model API to connect models to your application. Choose a model and configuration in the catalog, then authenticate with a personal PAYG key.

  1. Compare models and current prices: check protocol, capabilities, context and availability.
  2. Prepare an API key with personal models:chat permission.
  3. Make your first request using the selected configuration’s example.
  4. Handle streams and errors and retain request identifiers.
  5. Inspect usage and costs: delivery and settlement are separate states.

Choose an interface

NeedReference
OpenAI Chat Completions formatPOST /v1/chat/completions
Anthropic Messages formatPOST /v1/messages
Discover models and configurationsCatalog API
Image or video generationMedia API availability
Membership, Agent or compliance APIsMembership OAuth and SDK, other APIs

Protocol compatibility does not imply support for every vendor endpoint. Models, parameters and prices depend on the configuration. Being visible in the catalog does not mean a model is callable.

Pricing and accounts

The domestic service settles in CNY; USD is a reference equivalent. API Credits and membership balances are separate. Pending prices are not free usage, and vendor reference prices are not platform charges. Read the billing guide.

Find documentation

Tutorials explain the steps, API references describe requests and responses, and guides cover streaming, tools, limits, errors and billing. A machine-readable OpenAPI specification is also available.