Set Organization Model

(v1)

PUT

Sets one of the organization's default models by the type path segment: platform (powers Confident AI's own AI features) or simulation (simulates user turns in conversation simulations), applying to every project without an override of its own. The provider's credential must already be configured on the organization via the model credentials endpoint, and providers blocked by the model provider policy return 403. Setting CONFIDENT_AI requires no credential and clears the model name.

PUT/v1/organization/models/{type}

curl -X PUT "https://api.confident-ai.com/v1/organization/models/{type}" \
  -H "CONFIDENT_API_KEY: <ORGANIZATION-API-KEY>" \
  -H "Content-Type: application/json" \
  -d '{
  "provider": "GEMINI",
  "name": "gemini-2.0-flash",
  "maxConcurrency": 5
}'

200

{
  "success": true,
  "data": {
    "model": {
      "id": "cm4xqd2rp0002abcdef987654",
      "type": "PLATFORM",
      "provider": "GEMINI",
      "name": "gemini-2.0-flash",
      "maxConcurrency": 5,
      "maxInputTokens": null,
      "projectId": null,
      "organizationId": "c290fdd8-6a02-4056-b4d9-3459a4dcbd9e"
    }
  }
}

Headers

CONFIDENT_API_KEYstringRequired
The organization API key for your Confident AI organization.

Path parameters

typeenumRequired
Which of the organization's models to set.

Request body

providerenumRequired
The model provider to run the model on. CONFIDENT_AI requires no credential; every other provider requires its credential to be configured first via the model credentials endpoints. The CUSTOM provider cannot be configured through the public API.
Show 16 enum valuesHide 16 enum values
- CONFIDENT_AI
- OPEN_AI
- ANTHROPIC
- GEMINI
- X_AI
- DEEPSEEK
- MISTRAL
- PERPLEXITY
- BEDROCK
- VERTEX_AI
- AZURE
- PORTKEY
- OPEN_ROUTER
- TRUE_FOUNDRY
- LITE_LLM
- HUGGING_FACE
namestring
The model name to use, for example gpt-4o-mini. Omit to clear it; ignored for CONFIDENT_AI.
maxConcurrencyinteger
Maximum number of concurrent calls made to the model. Omit or pass null to clear it.
maxInputTokensinteger
Maximum number of input tokens sent to the model per call. Platform and simulation models only; rejected on the evaluation path. Omit or pass null to clear it.

Response

successboolean
Indicates if the request was successful
dataobject
Show 2 propertiesHide 2 properties
- modelobject
  Show 8 propertiesHide 8 properties
  - idstring
    Unique identifier of the model configuration
  - typeenum
    What the model is used for
    Show 4 enum valuesHide 4 enum values
    EVALUATION
    PLATFORM
    GENERATION
    SIMULATION
  - providerenum
    The model provider to run the model on. CONFIDENT_AI requires no credential; every other provider requires its credential to be configured first via the model credentials endpoints. The CUSTOM provider cannot be configured through the public API.
    Show 16 enum valuesHide 16 enum values
    CONFIDENT_AI
    OPEN_AI
    ANTHROPIC
    GEMINI
    X_AI
    DEEPSEEK
    MISTRAL
    PERPLEXITY
    BEDROCK
    VERTEX_AI
    AZURE
    PORTKEY
    OPEN_ROUTER
    TRUE_FOUNDRY
    LITE_LLM
    HUGGING_FACE
  - namestring
    The configured model name; null when the provider's default is used
  - maxConcurrencyinteger
    Maximum number of concurrent calls made to the model
  - maxInputTokensinteger
    Maximum number of input tokens sent to the model per call
  - projectIdstring
    Set when the model is configured on a project
  - organizationIdstring
    Set when the model is configured on the organization
- sourceenum
  Returned for project reads and for platform and simulation model updates; whether the model in effect comes from a project override or the organization default
  Show 2 enum valuesHide 2 enum values
  - project
  - organization

Built byConfident AI