gpt-4o-transcribe create transcription stream bug

[redpancho]

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

https://platform.openai.com/docs/api-reference/audio/createTranscription this page suggests that stream: true works for the 2 new models gpt-4o-mini-transcribe and gpt-4o-transcribe, however the behavior is not as expected on the page.

import fs from "fs";
import OpenAI from "openai";

const openai = new OpenAI();

const stream = await openai.audio.transcriptions.create({
  file: fs.createReadStream("audio.mp3"),
  model: "gpt-4o-mini-transcribe",
  stream: true,
});

for await (const event of stream) {
  console.log(event);
}

This is valid code. However, the "event" variable will output one letter at a time rather than what the API Reference suggests (a full JSON object)

Looking at the SDK the response does not look like it is being sent to ResponseStream correctly.

Thank you.

To Reproduce

    const stream = await openai.audio.transcriptions.create({
        file: fs.createReadStream("./audio.mp3"),
        model: "gpt-4o-transcribe",
        language: 'en',
        stream: true,
        response_format: 'json',
    });

    for await (const event of stream) {
            console.log(event);
    }

OS

Windows

Node version

20.12.0

Library version

v4.89.0

[Lucas-Levandoski]

Just gave it a shot and it is running just fine on version 4.92.1, even returning it as a json