What is the limit on audio length when using Gemini API to do ASR task?

Gemini API
1

Question is as title. The command is like this:

response = model.generate_content([prompt, audio_file])

When audio length bigger than a few minutes(maybe 10 mimutes?), a lot of transcription will be lost.
So, does the API have a audio length limit? and if so, what is the limit?

Thanks.

2

While Gemini can accept audio input, it isn’t necessarily the best solution for this problem.

You may want to consider something like the Google Speech to Text API, which does have models that are tuned for long input and transcription.