Persistent Platform Instability on AI Studio! Seeking Community Confirmation and Official Clarity

Google AI Studio
, , , ,
1

I’m posting this in the hope of getting both community input and an official response, because what I’m experiencing no longer feels like isolated bugs, it feels like sustained, platform-wide instability.

Before anything else, I want to acknowledge context: AI Studio and the broader Gemini stack went through a significant disruption in February. That was understandable at the time, and most of us were willing to be patient through the recovery period. However, we are now approximately three months past that event, and instead of seeing convergence toward stability, the volume and diversity of issues being reported across this forum appear to be increasing.

1. Billing and Quota

  • Paid-tier projects continuing to display “daily limit reached”
  • Billing status shown as “Unavailable” or inconsistent with Cloud Billing
  • Pro-plan accounts being rate-limited on the first request of a session
  • Quota exhaustion despite active paid subscriptions
  • Tier mismatches between Cloud Billing (Tier 2) and AI Studio (Tier 1)
  • Credits failing to activate; reduction in $300 Cloud Credit support
  • Family Plan Pro entitlements not propagating into AI Studio

2. API and Models

  • Recurring 403, 429, and 503 errors across regions and accounts
  • Failures linking AI Studio to the Gemini API
  • API key creation blocked by “Permission Denied” errors
  • Increased model unavailability, particularly for image models
  • Gemini 3.1 Pro Preview: response truncation with finishReason: STOP, citation inconsistencies, and thinking-mode bugs
  • Safety filter behavior change resulting in full-response erasure rather than graceful stopping
  • Perceived output-quality regression from 2.5 Pro to 3.1 Pro

3. Application and UI

  • “Failed to load app” error loops
  • App list not rendering for multiple users
  • Preview screens loading blank or remaining out of sync due to caching
  • Dropdown menus closing unexpectedly
  • Editor scroll issues on both web and mobile
  • General latency on routine operations

4. GitHub Integration

  • Widespread sync failures across projects
  • Authentication and login issues
  • Corruption of binary assets (e.g., images) on push
  • Inability to pull the latest repository state
  • Repository lock-in with no clear path to re-establish the connection

5. Deployment (Cloud Run / Firebase)

  • “Corrupted source archive” errors during publishing
  • Failures to create or update Cloud Run services
  • No option to select region or an existing service during deployment
  • Firebase authentication domain errors

6. Data Integrity and Project Loss

  • Reports of complete project loss with failed Drive backups
  • Missing restore points
  • Spontaneous file deletion attributed to OPFS I/O timeouts
  • Codebase corruption following project migration

7. Feature Regressions

  • Speech generation returning “No Audio Generated”
  • Nano Banana 2 / Pro: removal of PNG download and reduced output resolution
  • User-supplied API key flow no longer permitted in shared apps
  • Reduction in sharing-permission options
  • Live chat broken since the most recent update

Questions to the community and to the AI Studio team:

  1. To other users: are any of you currently experiencing a stable, predictable workflow on AI Studio? If so, it would be helpful to understand what your configuration looks like, because that data point is genuinely valuable here.

  2. To the team: are any of these issues formally acknowledged with target resolution timelines? A consolidated status page or pinned update would help significantly.

  3. Is the current state of the platform considered transitional i.e., expected to stabilize as ongoing changes complete or are these issues being treated as the steady state?

I’m not posting this to vent, I genuinely need to know whether to keep building on AI Studio or move workloads elsewhere. This post is not intended as criticism for its own sake. AI Studio is a platform many of us have built real work on, and the goal here is constructive: to surface the cumulative picture, confirm whether others are seeing the same, and request clear communication from the team.

Three months on from a major incident, foundational capabilities, billing accuracy, API reliability, source-control integration, and project durability should be dependable. For many users, they currently are not. That gap is worth naming, and worth addressing transparently.

If you’re running into any of these too, please reply, even a “+1 on billing” or “+1 on GitHub sync” helps establish how widespread this actually is.

2

Hello, I am experiencing the same issues. I was able to deploy apps, assign billing, and secret keys. No issue for a couple of months. This week, there is no deploy button, only a remix button. I had issues assigning billing and secret keys. Now I am experiencing timeouts and access issues, and I can not deploy new apps for our organization. Something broke it all for me. I have been trying to get help through support.. which pointed me here. I’m not sure what else to do since my leadership wants this all working by tomorrow.

3

Thank you for sharing this, your experience directly corroborates what the original post is describing, and the timing you mention (working fine for months, then multiple capabilities breaking in the same week) is an important signal. It points to a platform-side change rather than an account issue.

What you’re describing touches three of the categories above at once: deployment (Deploy button replaced by Remix), billing (assignment failures), and API access (secret keys, timeouts). That combination is showing up in several other threads too, so you are definitely not alone.

A couple of quick things that may help before tomorrow:

  • Check in Cloud Console whether your project’s billing link is still intact several users have found it silently broken after the recent updates.

To the AI Studio team: this is a clear example of the real cost of the current instability, a working production setup broken within a week, with support redirecting users to the forum and no clear path forward. A consolidated status update would genuinely help.

+1 on deployment, +1 on billing, +1 on API access.

Best of luck with the deadline, please post back with anything you find, it’ll help others in the same spot.