Skip to main content
POST
/
models
/
v2
/
openai
/
v1
/
responses
Responses
curl --request POST \
  --url https://api.bytez.com/models/v2/openai/v1/responses \
  --header 'Authorization: <authorization>' \
  --header 'Content-Type: application/json' \
  --data '
{
  "model": "<string>",
  "input": "<string>",
  "max_output_tokens": 256,
  "temperature": 0.7,
  "stream": false,
  "reasoning": {
    "effort": "medium",
    "summary": "none"
  },
  "tools": [
    {
      "type": "function",
      "function": {
        "name": "<string>",
        "description": "<string>",
        "parameters": {}
      }
    }
  ],
  "tool_choice": "auto",
  "metadata": {},
  "user": "<string>",
  "include": [
    "<string>"
  ],
  "top_logprobs": 123
}
'
{
  "id": "<string>",
  "object": "<string>",
  "created": 123,
  "model": "<string>",
  "output": [
    {
      "type": "message",
      "role": "<string>",
      "content": [
        {
          "type": "output_text",
          "text": "<string>"
        }
      ]
    }
  ],
  "output_text": "<string>",
  "usage": {}
}

Headers

​
Authorization
string
required

Token for authentication

Body

application/json
​
model
string
required

The ID of the model to run (e.g., anthropic/claude-opus-4-5, openai/gpt-5.1)

​
input
required

The input to the model. Can be a string (simple prompt) or an array of chat-style messages. For richer multimodal inputs, use the array form with multi-part content.

​
max_output_tokens
integer
default:256

Maximum number of tokens to generate (counts reasoning + visible output for reasoning models)

​
temperature
number
default:0.7

Sampling temperature

​
stream
boolean
default:false

Whether to stream SSE events

​
reasoning
object

Optional "thinking" controls for supported reasoning models.

​
tools
object[]

Optional tool definitions for function/tool calling

​
tool_choice

Tool selection behavior

Available options:
auto,
none,
required
​
metadata
object

Arbitrary key/value metadata to attach to the request

​
user
string

End-user identifier (if supported)

​
include
string[]

Optional list of extra fields to include in the response (e.g. logprobs). Example values may include message.output_text.logprobs (provider/model dependent).

​
top_logprobs
integer

Number of most likely tokens to return at each position (if logprobs are included)

Response

Successful response

​
id
string
required

Unique ID for this response

​
object
string
required

Type of returned object (usually response)

​
created
integer
required

Unix timestamp of response creation

​
model
string
required

Model used to generate the response

​
output
object[]

Output items (messages, reasoning summaries, tool calls, etc.)

​
output_text
string

Convenience field containing concatenated output text (when applicable)

​
usage
object

Token usage details (shape may vary by provider)