Batch API does not support image_size in image_config?

Google Cloud Build with AI Generative AI & Foundational Models
1

Using gemini-3.1-flash-image-preview , the image_size parameter works in the sync API but is rejected by the Batch API.

Sync API, works:

config = types.GenerateContentConfig(
    response_modalities=["IMAGE"],
    image_config=types.ImageConfig(
        image_size="512"  # Accepts "512", "1K", "2K", "4K"
    ),
)
response = client.models.generate_content(model="gemini-3.1-flash-image-preview", ...)

Batch API, fails:

JSONL uploaded to GCS:

{
  "key": "request-001",
  "request": {
    "contents": [{"role": "user", "parts": [{"text": "Generate a red circle"}]}],
    "generation_config": {
      "response_modalities": ["IMAGE"],
      "image_config": {"image_size": "512"}
    }
  }
}

Batch job created via:

client.batches.create(
    model="gemini-3.1-flash-image-preview",
    src="gs://bucket/request.jsonl",
    config=types.CreateBatchJobConfig(dest="gs://bucket/results/"),
)

Job completes but response contains error:

Failed to parse JSON into proto ... @ generation_config.image_config.image_size: string ... unexpected character: '5'; expected '"'

Is image_size supported in the Batch API?