Client.chat.completions.create hangs when using gpt-4o-audio-preview from AWS Lambda

API Bugs
1

When calling client.chat.completions.create from AWS Lambda using the new gpt-4o-audio-preview model this line just hangs forever (until the Lambda times out ~20s later):

completion = client.chat.completions.create(
    model=model,
    messages=all_msgs,
    modalities=['text'],
    seed=0,
    temperature=0,
)

I added a logging statement just before and just after and the first one gets printed and the second not.

What’s a bit puzzling is why this works when I call the same API with the same arguments from my local workstate.