Is the batch API very slow today for others?

Gemini API
, ,
1

Usually the Gemini Batch api completes within a couple hours for me but today I have the beliow

| Name | Requests | Age |

|-------------------------|------------|--------------|

| enrich-… | 9 pending | ~10 min |

| batch_0 | 50 pending | ~17 hours :warning: |

This is fairly unusual, one of the batches has gone on for 17+ hours, I’m on tier 1 (should be tier 2 soon but the button isn’t available yet) and it wasn’t a problem before today. This is on gemini-2.5-flash so not sure if the model is the cause I have yet to test on gemini-3

Anyone else experience unusual times for the batch API?

4 Likes
2

yes i do get quota problem plus it takes so long i am using gemini-3-flash

2 Likes
3

I am on Tier 1 and am using Gemini 2.5 Pro, and I encountered almost the same issue.
My task submitted via the Batch API remained in a pending state for over 24 hours, even though it is usually completed in less than 10 minutes.
I retried the task several times, but the behavior did not change.
However, the same task completed successfully and without issues when using the synchronous API.
This suggests that the issue may be specific to the Batch API.

1 Like
4

Yep, I think it’s only the Batch API with this issue I have tested with several of the models to see if it’s model specific, gemini-3-flash-preview and gemini-2.0-flash also seem to have the same problems.

Status on google systems looks all good but there is clearly an issue, SLA for batches are usually processed within 24 hours, only thing i can think of is a zombie failure where it’s not reporting it back and failing silently

3 Likes
6

Same happens with us from yesterday. Everything is in pending status for over 20 hours.

1 Like
8

Hey All,

Thanks for your report. We’ve pushed a fix that should resolve the problem. Please check and let us know if the problem persists.

1 Like
10

Hi @chanduriv it seems like there is an ongoing issue again with the Batch API. I have the following that have been pending for around 6 hours at BATCH_STATE_PENDING

Models: gemini-3-flash-preview

I have also seen another post with a possibly related issue [Batch API Error] "Deadline expired before operation could complete" or "The operation was cancelled" for gemini-3-flash-preview

11

Yeal, my batch hangs for 12 hours now.

…

[Batch Gemini] - batches/xxx- State: BATCH_STATE_PENDING - Progress: ?/34 (failed: 0) - Elapsed: 45511s

Previously, my batches hang less time but return these errors with very high rates [Batch API Error] "Deadline expired before operation could complete" or "The operation was cancelled" for gemini-3-flash-preview

12

Update after 17+ hours, Gemini-2.5-flash functions normally. Only Gemini 3 Flash preview is abnormal right now

13

Same here. Submitted a gemini-3-flash-preview batch coming up on 26 hours ago that is still in pending status. We heavily use cached content as well, deleting them upon batch completion. So these batches not completing is racking up significant cache storage costs for us as well :frowning:

1 Like
14

yes facing the same issue gemini 3 flashjob is processing for a long time

1 Like