How can I retrieve the token count for each OpenAI Assistant API call?

Below is the key information you need:

Token Consumption Data Location

The token usage details are not provided in the HTTP response headers but as part of the JSON response body. When you call the API (for example, when using chat completions or assistant endpoints), the response includes a field called usage. This object provides breakdowns like:

• prompt_tokens
• completion_tokens
• total_tokens

Example Response Snippet

JSONCopied
{
  "id": "chatcmpl-123456",
  "object": "chat.completion",
  "created": 1728933352,
  "model": "gpt-4o-2024-08-06",
  "choices": [
    {
      "index": 0,
      "message": {
        "role": "assistant",
        "content": "Hi there! How can I assist you today?",
        "refusal": null
      },
      "logprobs": null,
      "finish_reason": "stop"
    }
  ],
  "usage": {
    "prompt_tokens": 19,
    "completion_tokens": 10,
    "total_tokens": 29
  }
}

How to Use This Data

You can extract the token consumption information programmatically and use it to calculate charges for users. For each API call made:

1. Parse the JSON response.
2. Read the usage field.
3. Use fields like total_tokens to determine the cost based on your pricing model.

Additional Details

• Chat Completions: The token usage breakdown is typically present in every API response.
• Assistant File and Other Endpoints: Similar token statistics are available, sometimes showing additional details (such as for cached tokens or different token types).

By processing the usage object from the API responses, you can effectively track and bill for token consumption in your platform.