Models & parameters
List models and the Chat Completions fields you will use most
GET /v1/models lists everything your key can call. Treat the id field
as the value for model in a request — it is the only identifier the API accepts.
curl https://api.orientx.ai/v1/models \
-H "Authorization: Bearer $ORIENTX_API_KEY"{
"object": "list",
"data": [
{ "id": "<MODEL_ID>", "object": "model", "owned_by": "orientx" }
]
}Availability depends on your account, so query this endpoint rather than hardcoding a list. A model
that has been retired returns 404 with code model_not_found. The Quick Start
shows the same call in Python and TypeScript.
Request parameters
The Chat Completions endpoint accepts the standard OpenAI fields. The ones you will reach for most:
| Parameter | Type | Notes |
|---|---|---|
model | string | Required. An id from GET /v1/models. |
messages | array | Required. Ordered turns with role and content. |
stream | boolean | Return server-sent events instead of one response. |
temperature | number | Higher values produce more varied output. Defaults to 1. |
top_p | number | Nucleus sampling. Adjust this or temperature, not both. |
max_tokens | integer | Caps the completion length. Does not include the prompt. |
stop | string or array | Up to 4 sequences that end generation. |
Support for optional fields such as tools, response_format, and vision inputs varies by model.
When a model cannot honor a parameter, the API returns 400 naming the field rather than silently
ignoring it.
Streaming is covered in the Quick Start. Token counts on a
stream need stream_options: { "include_usage": true } — see usage & billing.