Completions

Generate text with the OpenAI Legacy Completions–compatible format.

POST /v1/completions

Legacy Completions takes a as input. It suits workloads that still use older OpenAI SDKs or simple text completion. For new chat scenarios, prefer Chat Completions.

The default example model is .

Available models

Model	Description
	Legacy Completions instruct 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
	string | array	Yes	Input prompt. May be a string, string array, token array, or array of token arrays; official default generates from the document start marker
	string	No	Text that appears after the completion insert point; only supported by
	integer	No	Maximum tokens to generate; official default ; prompt tokens plus this value must not exceed the model context length
	number	No	Sampling temperature, range –, default ; usually tune either or
	number	No	Nucleus sampling, range –, default
	integer	No	How many completions to generate per prompt, range –, default
	boolean	No	Whether to stream the response, default ; when enabled, pushes SSE and ends with
	object	No	Streaming options; only effective when
	integer	No	Number of most likely token log probabilities to return, maximum
	boolean	No	Whether to echo the prompt in addition to the completion, default
	string | array	No	Up to 4 stop sequences; generation stops when encountered
	number	No	Range –, default
	number	No	Range –, default
	integer	No	Server generates multiple candidates and returns the best one, default , maximum ; cannot be used with streaming; when used with , must be greater than
	object	No	Bias map for token appearance probabilities; keys are token IDs, values –
	integer	No	Random seed; best-effort deterministic sampling
	string	No	Unique end-user identifier for abuse monitoring

stream_options

Name	Type	Required	Description
	boolean	No	Push an extra chunk containing before

Request example

Basic completion

curl -X POST https://octopusx.ai/v1/completions \
  -H "Authorization: Bearer <token>" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-3.5-turbo-instruct",
    "prompt": "Rewrite the following sentence more formally: This API is pretty easy to use.",
    "max_tokens": 120,
    "temperature": 0.3
  }'

Streaming

curl -X POST https://octopusx.ai/v1/completions \
  -H "Authorization: Bearer <token>" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-3.5-turbo-instruct",
    "prompt": "Introduce the Completions endpoint in one sentence.",
    "max_tokens": 64,
    "stream": true,
    "stream_options": {
      "include_usage": true
    }
  }'

Response example

Non-streaming success

{
  "id": "cmpl_abc123",
  "object": "text_completion",
  "created": 1735689600,
  "model": "gpt-3.5-turbo-instruct",
  "choices": [
    {
      "index": 0,
      "text": "This API offers good usability.",
      "finish_reason": "stop"
    }
  ],
  "usage": {
    "prompt_tokens": 22,
    "completion_tokens": 10,
    "total_tokens": 32
  }
}

Streaming chunks (illustrative)

data: {"id":"cmpl_abc123","object":"text_completion","choices":[{"index":0,"text":"This","finish_reason":null}]}

data: {"id":"cmpl_abc123","object":"text_completion","choices":[{"index":0,"text":" API","finish_reason":null}]}

data: [DONE]

Response fields

Name	Type	Description
	string	Completion response ID
	string	Object type; always
	integer	Response creation timestamp (Unix seconds)
	string	Model that generated the response
	array	Array of candidate results
	integer	Candidate index
	string	Text generated by the model
	string	Finish reason, e.g.
	object	Token usage statistics
	integer	Input token count
	integer	Output token count
	integer	Total token count

Related endpoints

• Chat Completions
• Responses
• Text APIs overview