Batch job stuck in JOB_STATE_RUNNING for 32+ hours - Gemini 3 Pro Image

Gemini API
,
1

I submitted a batch job to generate ~2K images using gemini-3-pro-image-preview model. The job has been in JOB_STATE_RUNNING state for over 32 hours now, is it normal? What is strange is the last update time is too long ago

Job details:

  • Job ID: batches/12h57gulbhjv25zn4pw2nli7esk25wdld2ai

  • Model: gemini-3-pro-image-preview

  • Number of requests: 2,068 images

  • Created: Nov 24, 2025 16:41 UTC

  • Last update: Nov 24, 2025 19:11 UTC (no updates for 29+ hours)

  • Current state: JOB_STATE_RUNNING

  • No error messages

3

Hi @Dien_Hoa_Truong , Apologies for the delayed response.

Could you please confirm if you’re still facing the same issue? If so, please try with a Minimal Batch Job, if the small job processes quickly, the issue is likely related to the size of your original batches and the current resource availability.

5

In the end, it finished after 3 days, but returned me only half of my requests ( submit 2000 images, return only 1000 )

7

@Dien_Hoa_Truong, You can try to split your batch job requests in smaller parts, since smaller batches might optimize resource allocation.