More audio file type support in (openai-compatible) api?

Gemini API
,
1

i am trying to use audio in the openai compatible api:
however i get:
openai.BadRequestError: Error code: 400 - [{‘error’: {‘code’: 400, ‘message’: ‘Invalid audio format “m4a” for audio generation. Valid formats are: [wav, mp3]’, ‘status’: ‘INVALID_ARGUMENT’}}]

although this says more format are supported:

Supported audio formats

Gemini supports the following audio format MIME types:

  • WAV - audio/wav
  • MP3 - audio/mp3
  • AIFF - audio/aiff
  • AAC - audio/aac
  • OGG Vorbis - audio/ogg
  • FLAC - audio/flac

and this (Audio understanding (speech only)  |  Generative AI on Vertex AI  |  Google Cloud Documentation) says even more format are supported:

Audio MIME type Gemini 2.0 Flash Gemini 2.0 Flash-Lite
AAC - audio/aac
FLAC - audio/flac
MP3 - audio/mp3
MPA - audio/m4a
MPEG - audio/mpeg
MPGA - audio/mpga
MP4 - audio/mp4
OPUS - audio/opus
PCM - audio/pcm
WAV - audio/wav
WEBM - audio/webm

sorry for the quick typing. time constrained. but thanks in advance!

2

what puzzled me and prompted me to try was that this (m4a) works at ai studio…

3

Hi,

The OpenAI comp runs against the Google AI Gemini API and there I do not see the m4a format as an accepted option either.

Cheers

4

is this an excuse for not accepting m4a? Aren’t we passing the base64 encoded version and a string for media_type?

6

Hi @Sirui_Lu ,

Are you still facing this issue?