whisper-1 withdrawn: use gpt-4o-mini-transcribe

whisper-1 was withdrawn on October 28, 2026, and requests for it now return 503. Switch to the replacement model below: the endpoint and request format stay the same, but subtitle and timestamp formats are no longer available.

Published: October 28, 2026
Last checked: October 28, 2026
Status: Withdrawn

Key facts

Item	Details
Withdrawn model	`whisper-1`
Withdrawn on	October 28, 2026
What you see	HTTP 503 with `no available channels for model whisper-1`
Replacements	`gpt-4o-mini-transcribe` for everyday transcription, `gpt-4o-transcribe` when accuracy matters most
Endpoint	Unchanged: `POST https://api.laozhang.ai/v1/audio/transcriptions`
Response formats	The new models accept only `json` and `text` for `response_format`

Who is affected

Scope	Details
Affected	Transcription requests with `model` set to `whisper-1`, including requests for SRT or VTT subtitles or timestamps
Not affected	`gpt-4o-mini-transcribe`, `gpt-4o-transcribe`, and the speech models `tts-1`, `tts-1-hd`, and `gpt-4o-mini-tts`

How to migrate

Step	What to do
1	Change `model` from `whisper-1` to `gpt-4o-mini-transcribe`
2	If the request sets `response_format`, remove it or set it to `json` or `text`
3	Test with a short recording, check the transcript in the `text` field, then move all traffic

export LAOZHANG_API_KEY="YOUR_LAOZHANG_API_KEY"

curl --fail-with-body --max-time 300 \
  "https://api.laozhang.ai/v1/audio/transcriptions" \
  -H "Authorization: Bearer $LAOZHANG_API_KEY" \
  -F "model=gpt-4o-mini-transcribe" \
  -F "file=@audio.mp3"

Prices for both replacement models are in the model catalog, and the final charge is in your call logs. For language, prompt, and file-upload details, see Transcription and speech API.

What if I need subtitles or timestamps

gpt-4o-mini-transcribe and gpt-4o-transcribe cannot return SRT or VTT subtitles or verbose_json with timestamps; those values return an error. LaoZhang API does not offer a replacement model for this yet. If your workflow depends on subtitles or timestamps, contact the site owner or support and describe the use.

FAQ

What happens if I still call whisper-1?

The request returns HTTP 503 with a message that the model has no available channel. It is not routed to another model.

Will whisper-1 come back?

There is no timetable. If that changes, this page and the announcement index will be updated.

What do I need to change after switching to gpt-4o-mini-transcribe?

Usually only model. If you set response_format, remove it or set it to json or text. The endpoint, API key, and file upload stay the same.

Does it handle Chinese differently?

Both replacement models handle Chinese. Set language=zh and put names and product terms in prompt; see Set the language and add a prompt.

Which model is more accurate?

gpt-4o-transcribe, with the same one-line change to model. Its price is higher than gpt-4o-mini-transcribe; see the model catalog.

Transcription and speech API: parameters, examples, and FAQ for transcription and speech
OpenAI speech-to-text guide: supported audio formats and file size limits
Model catalog and call logs

​Key facts

​Who is affected

​How to migrate

​What if I need subtitles or timestamps

​FAQ

​What happens if I still call whisper-1?

​Will whisper-1 come back?

​What do I need to change after switching to gpt-4o-mini-transcribe?

​Does it handle Chinese differently?

​Which model is more accurate?

​Related documentation