Using OpenAI o1 and o3 Reasoning Models on KiosAPI

Use o1-mini, o1-preview, o3, o3-mini via POST /v1/chat/completions. Note: streaming is not supported and only user role messages are accepted.

KiosAPI supports OpenAI’s o1 and o3 reasoning model series through the standard POST /v1/chat/completions endpoint. These models are built for complex, multi-step reasoning tasks, but they come with two key behavioural differences from regular chat models: streaming is not supported, and the system role is not accepted — you must structure your conversation using user and assistant messages only.

Streaming is not supported for o1 and o3 series models. Always set stream: false (or omit it entirely) in your request body, otherwise the request will fail.

The system role is not supported. Only user and assistant roles are accepted in the messages array.

Request

POST https://kiosapi.com/v1/chat/completions

Headers

Header	Required	Description
Content-Type	Yes	application/json
Authorization	Yes	Bearer sk-xxx

Request Body

model · string · required

The reasoning model to use. Supported values include o1-mini, o1-preview, o3, and o3-mini.

messages · array · required

An array of message objects forming the conversation. Only user and assistant roles are supported.

Show messages[] properties

messages[].role · string · required

The role of the message author. Accepted values are "user" and "assistant". The "system" role is not supported by o1/o3 models.

messages[].content · string · required

The text content of the message.

max_completion_tokens · number

The maximum number of tokens the model is allowed to generate in its response, including reasoning tokens consumed internally.

stream · boolean

Whether to stream the response. Must be false — streaming is not supported for o1/o3 models.

Example Request

cURL

curl https://kiosapi.com/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer sk-xxx" \
  -d '{
    "model": "o1-mini",
    "messages": [
      {
        "role": "user",
        "content": "Explain quantum computing in simple terms"
      }
    ],
    "max_completion_tokens": 1688,
    "stream": false
  }'

Python

import openai

openai.api_key = "sk-xxx"
openai.base_url = "https://kiosapi.com/v1/"

response = openai.chat.completions.create(
    model="o1-mini",
    messages=[
        {"role": "user", "content": "Explain quantum computing in simple terms"}
    ],
    max_completion_tokens=1688,
    stream=False
)

print(response.choices[0].message.content)

Response

The response follows the standard chat completions shape. The finish_reason will be "stop" when the model has completed its reasoning and produced a final answer.

{
  "id": "chatcmpl-A1iMgDLzZtUJ9QDpfqDLxKH0zfUnp",
  "object": "chat.completion",
  "created": 1724972230,
  "model": "o1-mini",
  "choices": [
    {
      "index": 0,
      "message": {
        "role": "assistant",
        "content": "Quantum computing is a type of computation...",
        "refusal": null
      },
      "logprobs": null,
      "finish_reason": "stop"
    }
  ],
  "usage": {
    "prompt_tokens": 9,
    "completion_tokens": 150,
    "total_tokens": 159
  },
  "system_fingerprint": "fp_157b3831f5"
}

Response Fields

choices[].message.content · string

The model’s final text response after completing its internal reasoning process.

choices[].finish_reason · string

The reason the model stopped generating. Typically "stop" when a complete response has been produced.

usage.completion_tokens · number

Total completion tokens used, which may include tokens consumed during the model’s internal reasoning steps.