JSONKit
AI & LLM

OpenAI Chat Completion Response JSON Example

A real-world JSON example of an OpenAI Chat Completions API response — choices, message content, finish reason, and token usage. Copy-ready for building LLM integrations and cost tracking.

{
  "id": "chatcmpl-9kZpQ2rXvL8mNwT4hY6jK",
  "object": "chat.completion",
  "created": 1752134400,
  "model": "gpt-4o-2024-08-06",
  "choices": [
    {
      "index": 0,
      "message": {
        "role": "assistant",
        "content": "The three primary colors are red, blue, and yellow."
      },
      "finish_reason": "stop"
    }
  ],
  "usage": {
    "prompt_tokens": 18,
    "completion_tokens": 12,
    "total_tokens": 30
  },
  "system_fingerprint": "fp_a1b2c3d4e5"
}

Field Reference

idrequiredstringUnique identifier for this completion, prefixed chatcmpl- — useful for logging and support requests
choicesrequiredarrayOne entry per requested completion (n parameter) — usually a single-element array unless multiple completions were requested
choices[].finish_reasonrequiredstringWhy generation stopped: 'stop' (natural end), 'length' (hit max_tokens), 'tool_calls', or 'content_filter'
choices[].message.contentrequiredstring | nullThe generated text — null when the model instead returned a tool_calls array
usage.total_tokensrequirednumberprompt_tokens + completion_tokens — the number billing is based on
modelrequiredstringThe exact model version that served the request, useful since 'latest' aliases can change silently

Variants

Tool Callfinish_reason is 'tool_calls' instead of 'stop', and content is null — the model wants to call a function.
{
  "id": "chatcmpl-9kZqR3sYwM9nOx5iZ7kL",
  "object": "chat.completion",
  "created": 1752134500,
  "model": "gpt-4o-2024-08-06",
  "choices": [
    {
      "index": 0,
      "message": {
        "role": "assistant",
        "content": null,
        "tool_calls": [
          {
            "id": "call_8mKzQ2vXpL9wRfT",
            "type": "function",
            "function": {
              "name": "get_weather",
              "arguments": "{\"location\":\"Ahmedabad\",\"unit\":\"celsius\"}"
            }
          }
        ]
      },
      "finish_reason": "tool_calls"
    }
  ],
  "usage": {
    "prompt_tokens": 62,
    "completion_tokens": 21,
    "total_tokens": 83
  }
}

Common Use Cases

  • →Parsing the assistant's reply out of an LLM API integration
  • →Tracking per-request token usage for cost monitoring and rate limiting
  • →Building a mock LLM response for frontend development against a chat UI before wiring up the real API
openaillmchat completiongptAItokens

Validate or format this JSON

One click loads this exact example into the tool — no copy-paste needed. Format it, validate it, explore the tree, or generate TypeScript types instantly.

FormatValidateExplore treeGenerate TypeScriptMinify

Frequently Asked Questions

A finish_reason of 'length' means the response was truncated mid-thought because it hit the max_tokens limit — the content you got back may be an incomplete sentence or a cut-off JSON object, and you should either increase max_tokens or treat it as incomplete rather than a finished answer.

From usage.prompt_tokens and usage.completion_tokens separately, since input and output tokens are usually priced differently (output tokens typically cost more per token than input tokens).

In this specific example it can't — content is only null when the model is making a tool call (finish_reason: 'tool_calls'). A 'stop' finish_reason with null content would indicate a malformed or unexpected API response worth logging.

Related JSON Examples

API DesignREST API ResponseGeneric REST API response wrapper with data payload, status, and pagination meta.API DesignError ResponseREST API error JSON with status code, machine-readable error code, and validation details.API DesignPagination ResponsePaginated list response JSON with offset, cursor, total, and navigation links.