text_tokens and image_tokens not documented or typed

[NikkiAung]

Confirm this is a Node library issue and not an underlying OpenAI API issue

• This is an issue with the Node library

Describe the bug

When making a chat completions request to gpt-4o-audio-preview or gpt-4o-mini-audio-preview, the usage response contains text_tokens and image_tokens under prompt_tokens_details:

"prompt_tokens_details": {
      "cached_tokens": 0,
      "audio_tokens": 21,
      "text_tokens": 11,
      "image_tokens": 0
    },

And at openai-node/src/resources/completions.ts, image_tokens and text_tokens are not being defined. As a result

details = response.usage.prompt_tokens_details
details.text_ # ← No autocomplete!
details.text_tokens # ← Type checker error!

After my code changes

details = response.usage.prompt_tokens_details
details.text_ # ← IDE autocompletes to "text_tokens"! ✅
details.text_tokens # ← Type checker OK! ✅

To Reproduce

Make a request to chat completions API with model gpt-4o-audio-preview or gpt-4o-mini-audio-preview

Screen.Recording.2025-12-11.at.12.04.49.AM.mov

Code snippets To Reproduce Issue

import OpenAI from "openai";
import fs from 'fs/promises';
import path from 'path';
import { fileURLToPath } from 'url';
import { dirname } from 'path';
import dotenv from 'dotenv';

// Get the directory name of the current module
const __filename = fileURLToPath(import.meta.url);
const __dirname = dirname(__filename);

// Load environment variables from .env file in the script's directory
dotenv.config({ path: path.join(__dirname, '.env') });

const openai = new OpenAI({
  apiKey: process.env.OPENAI_API_KEY
});

async function analyzeAudio(audioFilePath) {
  try {
    // Read the audio file and convert it to a base64 string
    const audioBuffer = await fs.readFile(audioFilePath);
    const base64str = audioBuffer.toString('base64');

    const response = await openai.chat.completions.create({
      model: "gpt-4o-audio-preview",
      modalities: ["text", "audio"],
      audio: { voice: "alloy", format: "wav" },
      messages: [
        {
          role: "user",
          content: [
            { type: "text", text: "Describe this sound?" },
            { type: "input_audio", input_audio: { data: base64str, format: "wav" }}
          ]
        }
      ]
    });

    // Extract and print the transcript
    const transcript = response.choices[0].message.audio.transcript;
    console.log('Transcript:', transcript);
    
    // Log the full response for debugging
    console.log('\nFull Response:', JSON.stringify(response, null, 2));
    
    // Log token usage details
    // const usage = response.usage.prompt_tokens_details;
    // console.log('\nText Tokens:', usage.text_tokens);
    // console.log('Image Tokens:', usage.image_tokens);

    // response.usage.prompt_tokens_details

    // need to add type for text_tokens and image_tokens
    

  } catch (error) {
    console.error('Error:', error);
  }
}

// Use the BAK.wav file from the script's directory
const audioFilePath = path.join(__dirname, 'BAK.wav');
analyzeAudio(audioFilePath);

To get BAK.wav, download from kaggle audio datasethttps://www.kaggle.com/datasets/crischir/sample-wav-audio-files

OS

macOS

Node version

v24.11.1

Library version

2.9.0

[zanllan23-dev]

Ok tanks