The Pending Jobs were completed.
same here pending for over 72 hour
3 Likes
After 3 days I simply got a JOB_STATE_FAILED 
Still pending well past 48h, submitted in the last 2-3 days, tier 2, Gemini-2-5-pro
Hey folks, sorry about the issues and the delayed response here. Our team is actively working on fixing this. We’ll circle back with an update once this is resolved.
2 Likes
At least after almost 2 days:
JOB_STATE_PENDING 2026-04-02 11:49:23.510183+00:00
JOB_STATE_SUCCEEDED 2026-04-04 06:33:23.208376+00:00
1 Like
After 10 days, the batch jobs have been successful over the past 3 days. They take some time to JOB_STATE_SUCCEEDED, but are functioning as expected.
Am getting 503 error from the gemini api
This issue is now resolved. Do reach out if you see any batches stuck for over 24 hours.
Does Batch API work for anyone? I’m still getting JOB_STATE_EXPIRED after jobs pending for 24 hours.
After seeing this I guess google has introduced new way of processing, ‘flex’ as service tier, who has tried this?
48hrs and still pending.
1 Like
Same here, over 48 hours still in pending. Looking at my logs in ai studio though it looks like at least some of them have started being processed. I’m using the python package to monitor, but an http call to https://generativelanguage.googleapis.com/v1beta/batches/BATCH_ID?key=API_KEY also says that the jobs are pending.
1 Like
Unfortunately the jobs will expire
I’m still getting JOB_STATE_EXPIRED after jobs running for 24 hours. Not a single job completed successfully since March 27 for me
I tried Gemini Flex API h with gemini-3.1-flash-image-preview but getting “503, ‘message’: 'This model is currently experiencing high demand. Spikes in demand are usually temporary. Please try again later”
nothing works
1 Like
Flex lowers criticality and will lead to more errors (failure) during peak hours, see: https://ai.google.dev/gemini-api/docs/flex-inference
1 Like