[Bug] Extra inputs are not permitted error when using image_config in Batch API (Nano-Banana)

[hans2356]

Description of the bug:

When using the google-genai Python SDK to submit a batch job for image generation (using the gemini-3-pro-image-preview model), the SDK throws a Pydantic validation error if image_config is included in the generation_config.

According to the official documentation and Google Colab examples, image_config (containing aspect_ratio and image_size) should be supported. However, the SDK's Pydantic model seems to reject these fields as "extra inputs".

Actual vs expected behavior:

Python

from google import genai

client = genai.Client(api_key='YOUR_API_KEY')

Basic inline request structure following the docs

inline_requests = [
{
"contents": [{
"parts": [
{"text": "Transform this image into pixel art."},
{"inline_data": {"mime_type": "image/jpeg", "data": "BASE64_DATA..."}}
],
}],
"generation_config": {
"response_modalities": ["IMAGE"],
"image_config": {
"aspect_ratio": "16:9",
"image_size": "2K"
}
}
}
]

client.batches.create(
model="models/gemini-3-pro-image-preview",
src=inline_requests,
config={'display_name': 'test-job'}
)
Expected behavior The batch job should be created successfully without validation errors, as image_config is a documented feature for the Nano-Banana model.

Actual behavior (Error Message)

Plaintext

ValidationError: 1 validation error for BatchJobSource
inlined_requests.0.generation_config
Extra inputs are not permitted [type=extra_forbidden, input_value={'response_modalities': ['IMAGE'], 'image_config': {'aspect_ratio': '16:9', 'image_size': '2K'}}, input_type=dict]

Any other information you'd like to share?

No response

[Junjie-Chu]

requests_data = [
    {"key": "image_request_1", "request": {"contents": [{"parts": [{"text": "A big letter A surrounded by animals starting with the A letter"}]}],'generation_config': {'response_modalities': ['TEXT', 'IMAGE'], "temperature": 0.0, "image_config":{"aspect_ratio": "1:1", "image_size": "1K"}}}}, # Be careful, for inline request it would be 'config' instead of 'generation_config'
    {"key": "image_request_2", "request": {"contents": [{"parts": [{"text": "A big letter B surrounded by animals starting with the B letter"}]}],'generation_config': {'response_modalities': ['TEXT', 'IMAGE'], "temperature": 0.0, "image_config":{"aspect_ratio": "1:1", "image_size": "1K"}}}}
]

Above works for me. Not sure about your case.

[kkorpal]

Hi @hans2356, Thanks for reaching out!
I actually just tried reproducing this with the same parameters and it worked fine on my end!

A couple of things to check:

Make sure you're on the latest SDK version (pip install -U google-genai).

Double-check that the image_config is nested correctly inside the generation_config block.

If you're still seeing the error, could you share a snippet of the code you're using? That’ll help me spot if it's a specific formatting issue."

[github-actions]

Marking this issue as stale since it has been open for 14 days with no activity. This issue will be closed if no further activity occurs.