List model access providers with the Admin API

With a custom model-access policy on, list catalog providers and models (enabled flags plus per-model parameters) over GET https://api.cursor.com/teams/model-access/providers and GET https://api.cursor.com/teams/model-access/providers/:provider/models. Official reference: Admin API → Model access. These routes are in preview and may change before general availability.

Authenticate with a team Admin API key (Basic auth, key as username, empty password). Reads require models:read or admin:*. Generic read:* keys cannot call these routes. Rate limit: 20 requests per minute.

Both GETs return 409 when the team does not have a custom policy (state is unrestricted or legacy). Seed policy first: Set model access configuration with the Admin API.

Path segments use catalog ids such as anthropic and claude-opus-4-6. GET responses also include display names.

List every provider

curl -X GET https://api.cursor.com/teams/model-access/providers \
  -u YOUR_API_KEY:

Example response:

{
  "teamId": 7,
  "state": "custom",
  "providers": [
    {
      "id": "anthropic",
      "displayName": "Anthropic",
      "enabled": true,
      "models": [
        {
          "id": "claude-opus-4-6",
          "displayName": "Opus 4.6",
          "enabled": true,
          "parameters": [
            {
              "id": "fast",
              "displayName": "Fast",
              "supportedValues": ["false", "true"],
              "allowedValues": ["false", "true"],
              "configuredDefaultValue": null,
              "catalogDefaultValue": "true"
            }
          ]
        }
      ]
    },
    {
      "id": "openai",
      "displayName": "OpenAI",
      "enabled": true,
      "models": [
        {
          "id": "gpt-5.4",
          "displayName": "GPT-5.4",
          "enabled": true,
          "parameters": [
            {
              "id": "reasoning",
              "displayName": "Reasoning",
              "supportedValues": ["low", "medium", "high", "xhigh", "max"],
              "allowedValues": ["low", "medium", "high"],
              "configuredDefaultValue": "high",
              "catalogDefaultValue": "medium"
            }
          ]
        }
      ]
    }
  ]
}

Top-level field	Meaning
teamId	Integer team ID
state	Policy state (custom when this GET succeeds)
providers	Catalog providers with nested models

Provider field	Meaning
id	Catalog provider id (use this in path segments)
displayName	Human-readable label
enabled	Whether the provider is allowed for the team
models	Nested model objects

Model parameters fields

Each model includes a catalog-driven parameters array. Parameter ids and supported values come from the model catalog (for example fast, reasoning, effort, context). Use this GET to discover which parameters a model supports before writing.

Field	Meaning
id	Parameter id (for example fast or reasoning)
displayName	Human-readable label
supportedValues	All values the catalog allows for this parameter on this model
allowedValues	Values currently allowed by the team policy
configuredDefaultValue	Admin-pinned default, or null when unset
catalogDefaultValue	Catalog default for the parameter on this model

List models for one provider

curl -X GET https://api.cursor.com/teams/model-access/providers/anthropic/models \
  -u YOUR_API_KEY:

Path param	Rules
provider	Catalog provider id (for example anthropic)

Parameter fields on each model match the providers response.

Pitfalls

• Calling while state is unrestricted or legacy — both GETs return 409 until configuration seeds a custom policy.
• Putting display names in the path — use catalog ids from the response id fields.
• Skipping this GET before a model PUT — write only parameter ids and values the catalog lists under supportedValues.
• Treating preview response shapes as frozen — fields can shift before GA.