Skip to content

Models and pricing

Which models you can call, and what they cost. Pass a model's ID in the model field of your request.

List available models

The model catalogue changes as models are added, so the live list, not a table in these docs, is authoritative. Two ways to read it:

curl https://api.mangoboost.io/v1/models \
  -H "Authorization: Bearer $MANGOINFERENCE_API_KEY"

Returns the OpenAI-compatible model list: one entry per model, keyed by the id you pass as model. Only models your key can call are listed.

Sign in to the console at https://inference.mangoboost.io and open Models (https://inference.mangoboost.io/models).

Model IDs are the upstream repository IDs, including the vendor prefix (for example zai-org/GLM-5.3). Copy the ID exactly; it is case-sensitive.

Capabilities

Support for tool calling, structured output, and reasoning is a property of the model, not of the platform. A request using a capability the model was not trained for will either be ignored or fail. Check the model's own model card before relying on one, and test against your chosen model.

Pricing

Per-model prices are not published in these docs yet. Until they are, read your actual spend from the console (see Usage and cost breakdown) or ask for a rate sheet in Discord or at support@mangoboost.io.

What is settled about the billing mechanics:

  • Usage is metered in tokens. Every response reports what it consumed in its usage object. See Chat Completions.
  • Reasoning tokens are generated tokens. On a reasoning model they appear as usage.reasoning_tokens and are already counted inside completion_tokens, so they are billed like any other generated token. See Reasoning.
  • Usage draws down your credits/balance.