Create an AI model

POST /aisix/admin/models

Interactive request editor loads with JavaScript.

Query Parameters

ai_gateway_group_id*string

AI Gateway group ID (filter resources by group).

Length1 <= length

Request Body

application/json

name*string

Human-readable name for the model. Used as the model identifier in aisix proxy requests.

Match^[A-Za-z0-9][A-Za-z0-9._:/-]*$

Lengthlength <= 256

model*string

Underlying model identifier used by the provider.

Length1 <= length <= 256

provider_id*string

ID of the AIProvider that supplies credentials for this model. The provider must belong to the same AI Gateway group.

Length1 <= length <= 256

timeout?integer

Request timeout in milliseconds.

Formatint64

Range0 <= value

rate_limit?

Rate limit configuration for AI requests. Field names use the abbreviated forms that match aisix's internal representation (tpm/tpd/rpm/rpd/concurrency).

Response Body

`application/json`

curl -X POST "$API_BASE_URL/aisix/admin/models?ai_gateway_group_id=abc1234567890" \  -H "Content-Type: application/json" \  -d '{    "name": "my-gpt4o",    "model": "gpt-4o",    "provider_id": "provider-abc123"  }'

Successfully retrieved or modified the AI model.

{
  "value": {
    "id": "model-abc123",
    "ai_gateway_group_id": "group-abc123",
    "name": "openai/gpt-4o-alias",
    "model": "gpt-4o",
    "provider_id": "provider-abc123",
    "timeout": 30000,
    "rate_limit": {
      "tpm": 1,
      "tpd": 1,
      "rpm": 1,
      "rpd": 1,
      "concurrency": 1
    },
    "created_at": "2019-08-24T14:15:22Z",
    "updated_at": "2019-08-24T14:15:22Z"
  }
}

Bad Request — The request was malformed or contained invalid parameters. Check the error_msg field for details.

{
  "value": {},
  "error_msg": "error message"
}

Conflict — A resource with the same unique identifier (e.g., name) already exists.

{
  "value": {},
  "error_msg": "error message"
}

Complete operation details and schema variants

IAM Action: ai_gateway:CreateAIModel, Resource: arn:api7:ai_gateway:aigatewaygroup/%s Creates a new AI model configuration in the specified AI Gateway group. The model will be synced to the aisix data plane via etcd.

Parameters

ai_gateway_group_id (query, string, required): AI Gateway group ID (filter resources by group).

Request body

Content type: application/json

name (string, required): Human-readable name for the model. Used as the model identifier in aisix proxy requests.
model (string, required): Underlying model identifier used by the provider.
provider_id (string, required): ID of the AIProvider that supplies credentials for this model. The provider must belong to the same AI Gateway group.
timeout (integer, optional): Request timeout in milliseconds.
rate_limit (object, optional): Rate limit configuration for AI requests. Field names use the abbreviated forms that match aisix's internal representation (tpm/tpd/rpm/rpd/concurrency).
rate_limit.tpm (integer, optional): Maximum tokens allowed per minute (token_per_minute).
rate_limit.tpd (integer, optional): Maximum tokens allowed per day (token_per_day).
rate_limit.rpm (integer, optional): Maximum requests allowed per minute (request_per_minute).
rate_limit.rpd (integer, optional): Maximum requests allowed per day (request_per_day).
rate_limit.concurrency (integer, optional): Maximum concurrent requests allowed.

Responses

200: Successfully retrieved or modified the AI model.
allOf variant 1
value (object, optional): The specific resource.
allOf variant 2
value (object, optional): LLM model configuration in an AI Gateway group.
value.id (string, optional): Unique model ID.
value.ai_gateway_group_id (string, optional): ID of the AI Gateway group this model belongs to (control-plane only).
value.name (string, optional): Human-readable name for the model. Used as the model identifier in aisix proxy requests.
value.model (string, optional): Underlying model identifier used by the provider.
value.provider_id (string, optional): ID of the AIProvider entry that supplies credentials for this model. The provider must belong to the same AI Gateway group.
value.timeout (integer, optional): Request timeout in milliseconds.
value.rate_limit (object, optional): Rate limit configuration for AI requests. Field names use the abbreviated forms that match aisix's internal representation (tpm/tpd/rpm/rpd/concurrency).
value.rate_limit.tpm (integer, optional): Maximum tokens allowed per minute (token_per_minute).
value.rate_limit.tpd (integer, optional): Maximum tokens allowed per day (token_per_day).
value.rate_limit.rpm (integer, optional): Maximum requests allowed per minute (request_per_minute).
value.rate_limit.rpd (integer, optional): Maximum requests allowed per day (request_per_day).
value.rate_limit.concurrency (integer, optional): Maximum concurrent requests allowed.
value.created_at (string, optional): Timestamp when the model was created.
value.updated_at (string, optional): Timestamp when the model was last updated.
400: Bad Request — The request was malformed or contained invalid parameters. Check the error_msg field for details.
value (object, optional): The specific resource.
error_msg (string, optional): The error message.
409: Conflict — A resource with the same unique identifier (e.g., name) already exists. Response fields (application/json): identical to POST /aisix/admin/models, response 400 above.

cURL

curl -X POST '$API_BASE_URL/aisix/admin/models'

Create an AI model

200application/json

400application/json

409application/json

`application/json`

`application/json`

`application/json`