[Bug] Omni 1.1 video generation fails on Agent Platform with "AgentPlatform quota exhausted" while Omni 1.0 previously worked

Gemini API
1

Hello,

I would like to report a possible issue with Omni 1.1 video generation.

Before the Omni 1.1 update, I was able to successfully generate videos through Gemini Enterprise Agent Platform using Omni 1.0.

However, since Omni 1.1 was released, I have not had a single successful video generation through Agent Platform.

Every attempt fails with the following error:

“AgentPlatform quota exhausted”

My Google Cloud free trial is still active, with approximately 40,000 JPY worth of trial credits remaining. The trial is also close to expiration, with approximately 2 days remaining.

I also tested Omni 1.1 video generation through Google AI Studio.

To use the API project through AI Studio, I enabled billing by adding approximately 1,000 JPY.

After enabling billing, Omni 1.1 video generation through AI Studio succeeded once.

The usage for that successful generation appeared to be deducted from my remaining Google Cloud free trial credits.

However, after that single successful generation, AI Studio video generation also stopped working.

When attempting to generate a video, the AI Studio page fails during generation and needs to be reloaded.

At the same time, video generation continues to work normally through other Google interfaces:

  • Gemini Web: Working normally
  • Google Flow: Working normally

So the current situation is:

  • Agent Platform + Omni 1.0: Previously worked
  • Agent Platform + Omni 1.1: Never succeeded; consistently returns “AgentPlatform quota exhausted”
  • Google AI Studio + Omni 1.1: Succeeded once, then stopped working and the page fails during generation
  • Gemini Web: Working normally
  • Google Flow: Working normally

Because Agent Platform worked with Omni 1.0 but has not succeeded even once with Omni 1.1, while Gemini Web and Google Flow continue to generate videos normally, I am wondering if there may be an issue related to quota provisioning, entitlement, billing/trial credit routing, or backend configuration for Omni 1.1 on the API / Agent Platform side.

Could you please confirm whether this is a known issue, or advise what I should check on my project/account side?

I can provide screenshots, timestamps, model names, or additional error details if needed.

Thank you.

2

Hello,

Agent Platform related issues should be reported to Google Cloud Support or please open a topic on the Google Cloud Community Forum.

For the video generation issue in Google AI Studio, you mentioned that you were able to create the video once but never again, can you share more about that? what errors did you get after that? or if it was UI issue, if you could share screen recordings, screenshots etc, that’d be helpful. thanks!

3

Hello,

Thank you for your reply.

Regarding the Agent Platform issue, I understand. I will look into that separately through the appropriate support channel or community forum.

Regarding the video generation issue in Google AI Studio, I continued testing after my original report and found a more specific pattern.

When I start a new chat, the first video generation usually works normally. However, when I request a second, unrelated video within the same chat, the previous video generation sometimes incorrectly affects the new request.

For example:

  1. I start a new chat and generate the first video. The result is normal.

  2. I then enter a prompt for a completely separate and unrelated video in the same chat.

  3. The second video may unexpectedly become a continuation of the first video, or elements from the first video may be mixed into the new video, even though the second prompt does not ask to continue, modify, or reference the first video.

  4. If I start another new chat, the first video generation again usually works normally.

Because of this, creating a new chat appears to be a workaround, but I would not expect it to be necessary to start a completely new chat for every independent video generation. Previous video content should not affect an unrelated new request simply because it is generated within the same chat.

Regarding my original report that video generation succeeded once and then stopped working, further testing showed that the issue is not a complete inability to generate videos. Instead, starting a new chat appears to significantly improve reliability, while generating multiple independent videos within the same chat can cause previous video content to carry over into subsequent generations.

I can provide the prompts used for both videos, screenshots, screen recordings, and examples of the generated videos if that would be helpful for investigating the issue.

Thank you.