Create Responses

POST

/v1/responses

Generate responses from a large language model based on a given input. The Responses API provides a unified interface for text generation, multi-turn continuation, tool calling, reasoning configuration, and optional streaming.

Authorization

The AccessToken must be included in the request as a header when making REST API requests, along with the Content-Type header. You can use the following format for authorization:

--header 'Authorization: Bearer <your_token_here>'
--header 'Content-Type: application/json'

Note: Replace your_token_here with your actual AccessToken. It contains information that allows the server to verify your identity and permissions. You can create your API key here.

Request Body

Field	Type	Required	Description
background	boolean, null	No	Whether to run the model response in the background.
include	string[]	No	An array of additional output data to include in the response, such as file_search_call.results, message.output_text.logprobs, web_search_call.action.sources, or reasoning.encrypted_content.
model	string	Yes	The name of the model to use for generating responses
input	string, string[]	Yes	The input to use for generating responses, it can be a single string or an array of strings
instructions	string	No	Optional instructions to guide the model's response generation, providing specific directives or constraints for the output.
stream	boolean, null	No	Whether to stream the response back as it's generated (default: false)
max_output_tokens	integer, null	No	The maximum number of tokens to generate in the response
max_tool_calls	integer, null	No	The maximum number of tool calls allowed during response generation, helping to control the extent of external tool usage.
temperature	number, null	No	The sampling temperature to use when generating response
top_p	number, null	No	The nucleus sampling probability to use when generating response, helping balance randomness and coherence in the generated response.
text	object	No	Configuration options for text output, including plain text and structured JSON output through text.format.
parallel_tool_calls	boolean, null	No	Whether to allow parallel tool calls during response generation, enabling multiple tools to be called simultaneously for more efficient processing.
previous_response_id	string, null	No	The ID of a previous response to use as context for generating the new response, allowing for continued conversations or follow-up responses based on prior interactions.
metadata	object	No	An object containing additional metadata to include in the request, which can be used for various purposes such as tracking, logging, or providing extra context for response generation.
reasoning	object	No	An object containing configuration for the reasoning model, which can be used to enable or customize the reasoning capabilities of the model during response generation.
store	boolean, null	No	Whether to store the generated response in the system for future reference or analysis (default: false)
tool_choice	string, null	No	The strategy to use when the model needs to choose between multiple tools during response generation, such as auto, none, or required.
tools	array, null	No	An array of tool definitions that the model can call during response generation, allowing for enhanced functionality and integration with external systems. Each tool definition includes the tool's name, description, parameters, and other relevant information to guide the model in using the tool effectively.
truncation	string, null	No	The truncation strategy to use for the model response. Supported values are auto and disabled (default).

Tool Configuration

Currently, the API supports only function tools. Each object in the tools array should have the following structure:

Field	Type	Required	Description
type	string	Yes	The type of the tool. Currently, function is supported.
name	string	Yes	The function name that the model can call.
description	string, null	No	A description of what the function does, used by the model to decide when and how to call it.
parameters	object	Yes	A JSON Schema object that defines the function arguments.
strict	boolean, null	No	Whether the model should strictly follow the provided parameter schema when calling the function.

Example JSON Schema for parameters:

{
  "type": "object",
  "properties": {
    "location": {
      "type": "string",
      "description": "The city and country, for example: Singapore, Singapore"
    }
  },
  "required": ["location"],
  "additionalProperties": false
}

Reasoning Configuration

Field	Type	Required	Description
effort	string	No	The reasoning effort to use for reasoning-capable models, such as minimal, low, medium, or high.
summary	string, null	No	Controls whether reasoning summaries are generated when supported by the selected model.

Text Configuration

The text object configures the text output format and verbosity.

Field	Type	Required	Description
format	object	No	The response format. The default is { "type": "text" }. Use { "type": "json_schema" } for structured outputs, or { "type": "json_object" } for JSON mode.
verbosity	string, null	No	Controls output verbosity. Supported values are low, medium, and high.

The format object should have the following structure:

Field	Type	Required	Description
type	string	Yes	The format type. Supported values are text, json_object, and json_schema.
name	string	No	The name of the JSON Schema response format. Required when type is json_schema.
description	string	No	A description of the response format, used by the model to determine how to respond.
schema	object	No	The JSON Schema object that the model output must follow. Required when type is json_schema.
strict	boolean, null	No	Whether to enable strict schema adherence when generating structured output.

Include Configuration

The include array controls optional fields returned in the response. Common values include:

Value	Description
file_search_call.results	Include file search results.
message.output_text.logprobs	Include log probabilities for output text.
web_search_call.action.sources	Include sources from web search tool calls.
reasoning.encrypted_content	Include encrypted reasoning content when supported.

Example Request

{
  "model": "minimax/minimax-m2.5",
  "input": "Explain the concept of a polymer in simple terms.",
  "instructions": "Answer clearly and concisely.",
  "stream": false,
  "max_output_tokens": 100,
  "temperature": 0.7,
  "top_p": 0.9,
  "text": {
    "format": {
      "type": "text"
    }
  },
  "parallel_tool_calls": true,
  "store": false,
  "truncation": "disabled"
}

Response

Success Response

Field	Type	Description
id	string	The unique identifier for the response
object	string	The type of object returned, which is response for this endpoint
created_at	integer	The timestamp (in seconds since the Unix epoch) when the response was created
status	string	The status of the response, such as completed, in_progress, failed, or incomplete
completed_at	integer, null	The timestamp (in seconds since the Unix epoch) when the response completed, if available.
model	string	The name of the model used to generate the response
output	array	An array of output items generated by the model, such as messages, reasoning items, and tool calls.
output_text	string	SDK-only convenience field containing aggregated generated text when available. It is not included in the raw REST response body.
error	object, null	Error details if the response failed.
incomplete_details	object, null	Details explaining why the response is incomplete, if applicable.
instructions	string, null	The instructions used for this response.
max_output_tokens	integer, null	The maximum number of output tokens configured for the response.
max_tool_calls	integer, null	The maximum number of tool calls configured for the response.
parallel_tool_calls	boolean	Whether parallel tool calls were enabled for the response.
previous_response_id	string, null	The previous response ID used as context, if provided.
reasoning	object	The reasoning configuration used for the response.
store	boolean	Whether the response was stored.
temperature	number, null	The sampling temperature used for generation.
text	object	The text configuration used for generation.
tool_choice	string, object	The tool selection strategy used for generation.
tools	array	The tool definitions available to the model.
top_p	number, null	The nucleus sampling probability used for generation.
truncation	string, null	The truncation strategy used for the response.
usage	object	Token usage for the request and response.
metadata	object	Additional metadata associated with the response.

Create Responses

curl 'https://api.hpc-ai.com/inference/v1/responses'   -H 'Content-Type: application/json'   -H 'Authorization: Bearer <your_token_here>'   -d '{    "model": "minimax/minimax-m2.5",    "input": "Explain the concept of a polymer in simple terms.",    "instructions": "Answer clearly and concisely.",    "max_output_tokens": 100,    "temperature": 0.7,    "top_p": 0.9,    "text": {      "format": { "type": "text" }    },    "stream": false,    "store": false,    "truncation": "disabled"  }'  

{  "id": "resp_e63095aef9bc4d7292b769edb2cb6583",  "object": "response",  "created_at": 1773651537,  "status": "completed",  "completed_at": 1773651538,  "model": "minimax/minimax-m2.5",  "output": [    {      "type": "message",      "id": "msg_001",      "status": "completed",      "role": "assistant",      "content": [        {          "type": "output_text",          "text": "A polymer is a large molecule made by linking many smaller repeating units together, like beads on a string. Plastics, rubber, and DNA are all examples of polymers.",          "annotations": []        }      ]    }  ],  "error": null,  "incomplete_details": null,  "instructions": "Answer clearly and concisely.",  "max_output_tokens": 100,  "max_tool_calls": null,  "parallel_tool_calls": true,  "previous_response_id": null,  "reasoning": {    "effort": null,    "summary": null  },  "store": false,  "temperature": 0.7,  "text": {    "format": {      "type": "text"    }  },  "tool_choice": "auto",  "tools": [],  "top_p": 0.9,  "truncation": "disabled",  "usage": {    "input_tokens": 15,    "input_tokens_details": {      "cached_tokens": 0    },    "output_tokens": 34,    "output_tokens_details": {      "reasoning_tokens": 0    },    "total_tokens": 49  },  "user": null,  "metadata": {}}