Does using inline-requests in batch-mode guarantee the order of outputs?

For example, if I send a request through a command

inline_requests = []

for audio_path in audio_paths:
    with open(audio_path, 'rb') as f:
        audio_bytes = f.read()
    data = types.Part.from_bytes(data=audio_bytes, mime_type='audio/flac')
    inline_requests.append({'contents': [prompt, data]})
inline_batch_job = client.batches.create(
    model="models/gemini-2.5-flash",
    src=inline_requests,
    config={
        'display_name': "inlined-requests-job-1",
    },
)

In this case, will there be results for audio_paths[0] first, then for audio_paths[1], and so on?

Hi @Andrey_Libman
No, using inline requests in batch mode does not guarantee the order of outputs. The results of the batch job can be returned in any order.

but then how can I figure out which response relates to which query?
You are wrong I Think:

It’s comment from official tutorial: Google Colab

It says that, but it behaves differnetly. I am receiving results in scrambled order.

Exactly. I’m having the same issue so I made a bug report for it: [BUG] Batch API InlinedResponses are not returned in order · Issue #1909 · googleapis/python-genai · GitHub