Chat Completions

Call GPT models with the OpenAI Chat Completions–compatible format. Supports a single complete response or SSE streaming. See Available models below for options.

POST /v1/chat/completions

Use the unified chat message format to call GPT and related upstream models. stream controls how results are returned: omit it or set false for a single complete chat.completion object (good for background jobs, structured output, and short Q&A); set true for SSE incremental chunks (good for live chat, agents, and long-form generation).

When stream is omitted, the request is treated as non-streaming. To require streaming, pass "stream": true explicitly.

The default model is .

Available models

Model	Description
	Flagship latest in the 5.6 series
	High-spec 5.5 series
	General-purpose 5.4 series
	Lightweight 5.4
	Code-focused 5.3
	Smaller 5-series variant
	General multimodal model
	General text / tool-calling model

Headers

Name	Type	Required	Description
	string	Yes	Request authentication. Use a Bearer token, for example:
	string	Yes	Request content type; must be

Request body

Name	Type	Required	Description
	string	Yes	Model ID; default example . See Available models above
	array	Yes	Conversation messages in chronological order
	boolean	No	Whether to stream the response; defaults to non-streaming
	object	No	Streaming options; only effective when
	integer	No	Upper bound on generated tokens (visible output and reasoning tokens); recommended for reasoning models
	integer	No	Maximum generated tokens; deprecated, prefer
	number	No	Sampling temperature, range –, default
	number	No	Nucleus sampling, range –, default ; usually tune either or
	integer	No	Number of candidate results, range –, default
	string | array	No	Up to 4 stop sequences
	number	No	Range –, default
	number	No	Range –, default
	object	No	Bias map for token appearance probabilities
	boolean	No	Whether to return log probabilities for output tokens, default
	integer	No	–; requires to be
	integer	No	Random seed (Beta); best-effort deterministic sampling
	object	No	Output format; commonly used for JSON / structured output
	array	No	List of tools the model may call
	string | object	No	Controls tool calling: / / , or a specific function
	boolean	No	Whether to allow parallel tool calls, default
	string	No	Reasoning effort: / / / / / /
	string	No	Reply verbosity: / / ; supported by the GPT-5 series
	array	No	Output modalities, default ; audio models may include
	object	No	Audio output parameters; required when includes
	object	No	Predicted outputs configuration; can reduce latency for predictable content
	boolean	No	Whether to store this output, default
	object	No	Up to 16 key-value pairs of additional metadata
	string	No	Processing tier: / / / / /
	object	No	Web search options (search-capable models)
	string	No	Prompt cache key
	string	No	Stable end-user identifier, max 64 characters
	string	No	End-user identifier; deprecated, use /

messages items

Name	Type	Required	Description
	string	Yes	, , , , or
	string | array | null	Yes	Plain text string, or a multimodal part array ( / / / , etc.); for tool calls, content may be
	string	No	Optional participant name
	array	No	only: list of tool calls initiated by the model
	string	Conditional	Required for messages: corresponding tool call ID

stream_options

Name	Type	Required	Description
	boolean	No	Push an extra chunk containing before
	boolean	No	Stream obfuscation; normalizes payload size

response_format

Name	Type	Required	Description
	string	No	(default), , or
	object	Conditional	Provide when is ; includes , , etc.

tools items

Name	Type	Required	Description
	string	Yes	For function tools, always
	object	Yes	Includes , , (JSON Schema), and optional

Request example

Non-streaming

curl -X POST https://octopusx.ai/v1/chat/completions \
  -H "Authorization: Bearer <token>" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-5.6-terra",
    "messages": [
      { "role": "system", "content": "You are an API documentation assistant." },
      { "role": "user", "content": "Generate a short endpoint summary." }
    ],
    "stream": false
  }'

Streaming

curl -X POST https://octopusx.ai/v1/chat/completions \
  -H "Authorization: Bearer <token>" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-5.6-terra",
    "messages": [
      { "role": "user", "content": "Introduce Chat Completions in one sentence." }
    ],
    "stream": true,
    "stream_options": {
      "include_usage": true
    }
  }'

Multimodal

curl -X POST https://octopusx.ai/v1/chat/completions \
  -H "Authorization: Bearer <token>" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-5.6-terra",
    "messages": [
      {
        "role": "user",
        "content": [
          { "type": "text", "text": "Describe this image." },
          {
            "type": "image_url",
            "image_url": { "url": "https://example.com/image.png" }
          }
        ]
      }
    ]
  }'

Structured output

curl -X POST https://octopusx.ai/v1/chat/completions \
  -H "Authorization: Bearer <token>" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-5.6-terra",
    "messages": [
      { "role": "user", "content": "Extract the contact name and email." }
    ],
    "response_format": {
      "type": "json_schema",
      "json_schema": {
        "name": "contact",
        "strict": true,
        "schema": {
          "type": "object",
          "properties": {
            "name": { "type": "string" },
            "email": { "type": "string" }
          },
          "required": ["name", "email"],
          "additionalProperties": false
        }
      }
    }
  }'

Response example

Non-streaming success

{
  "id": "chatcmpl_abc123",
  "object": "chat.completion",
  "created": 1735689600,
  "model": "gpt-5.6-terra",
  "choices": [
    {
      "index": 0,
      "message": {
        "role": "assistant",
        "content": "This endpoint accepts unified chat messages and returns a single complete reply from the model."
      },
      "finish_reason": "stop"
    }
  ],
  "usage": {
    "prompt_tokens": 31,
    "completion_tokens": 24,
    "total_tokens": 55
  }
}

Streaming chunks (illustrative)

data: {"id":"chatcmpl_abc123","object":"chat.completion.chunk","choices":[{"index":0,"delta":{"role":"assistant","content":"This"},"finish_reason":null}]}

data: {"id":"chatcmpl_abc123","object":"chat.completion.chunk","choices":[{"index":0,"delta":{"content":" endpoint"},"finish_reason":null}]}

data: [DONE]

Response fields

Name	Type	Description
	string	Response ID for this generation
	string	for non-streaming; for streaming
	integer	Creation timestamp (Unix seconds)
	string	Model ID actually used
	array	Array of candidate results
	object	Non-streaming: message generated by the model
	string | null	Text content; may be when tool calls are present
	array	Function tools the model requested to call
	object	Streaming: incremental content ( / / / )
	string	Finish reason: , , , etc.
	object	Token usage; for streaming, depends on upstream support and often
	integer	Input token count
	integer	Output token count
	integer	Total token count

Notes

TIP

Parameters marked as optional on this page are not necessarily supported by every model. Whether a field is forwarded and takes effect depends on the capabilities of the model that ultimately handles the request.

WARNING

Streaming responses use text/event-stream. Clients must parse SSE data: lines and treat data: [DONE] as the end marker.

Related endpoints

• Responses
• Text APIs overview
• Error codes