After the very latest update, I observed that the models have become totally useless. They are not even responding back. Even on Pro account… ![]()
For Ultra the same,
in most cases it’s an issue with their servers and their capacity to handle traffic from all the subscribers
Even Ultra
!!?
hopefully, they get it fixed soon.
I am experiencing the same issue. I foolishly inquired with the agents as to why it had become so sluggish, proceeded to clear out the “unnecessary” files and folders they instructed me to, and even went so far as to uninstall and reinstall the program twice. However, I now realize that the problem does not stem from my settings or computer.
Hey All,
Thank you for bringing this to our attention. To help us troubleshoot, could you please share following details:
- The exact error message (e.g.,
Agent execution terminatedorModel overloadedorModel Unavailable)? - Which specific model is hanging (Gemini 3.1 Pro High/Low, Gemini 3 Flash, Claude Opus 4.6 or Claude Sonnet 4.6 )?
- Which region are you located in?
- Diagnostic Logs: If available, please share any logs from the terminal or output window
I’ve been having the same issue for a few weeks now. I’ve tried everything: reinstalling, clearing data, and I’ve even switched PCs in the meantime. But nothing worked. I have a Google AI subscription, and my projects aren’t huge; in fact, it’s significantly faster on other models like Gemini Flash or Claude.
There was no error message — it just kept showing ‘Working, Generating, Loading’ endlessly on Gemini 3.1 Pro (High) in India.
Claude models are working fast and fine.
Yeah, I’ve tried different nodes as well. Still no luck on my end too, but Claude works fine in Antigravity compared to Gemini models. Let’s hope it resolves.
Super slow response times after 3.1 pro. When asking something simple then it takes several minutes to respond and when using Claude code then limits are full after few hours and fast is usable and fast, but responses are way off.
Trajectory ID: bf8dd19d-360
Error: HTTP 429 Too Many Requests
Sherlog:
TraceID: 0x8ed8ceb1
Headers: {“Alt-Svc”:[“h3=“:443”; ma=2592000,h3-29=“:443”; ma=2592000”],“Content-Length”:[“139”],“Content-Type”:[“text/event-stream”],“Date”:[“Tue, 03 Mar 2026 09:06:37 GMT”],“Server”:[“ESF”],“Server-Timing”:[“gfet4t7; dur=80”],“Vary”:[“Origin”,“X-Origin”,“Referer”],“X-Cloudaicompanion-Trace-Id”:[“8ed8ceb191662026”],“X-Content-Type-Options”:[“nosniff”],“X-Frame-Options”:[“SAMEORIGIN”],“X-Xss-Protection”:[“0”]}
{
“error”: {
“code”: 429,
“message”: “Resource has been exhausted (e.g. check quota).”,
“status”: “RESOURCE_EXHAUSTED”
}
}
I was using antigravity almost from day one and I am paid user almost from start , but in last week its very unstable and slow and stop working too much
Experiencing the same issues since Friday, Feb 27th. Slow and eternal loops. Basically unusable at the moment
. Very frustrating! Google, please HELP!
Hi @KullarKert,
This error message indicates that you are sending too many requests, and the system is rate-limiting you. Even if you are within your overall daily quota, sending a very large number of requests in a short period can temporarily trigger this burst rate limit.
When you experience ‘super slow’ response times taking several minutes, it is likely because the tool you are using is hitting this rate limit and quietly retrying in the background (exponential backoff) before finally giving up and throwing the error you see here. If you are performing actions automatically or very rapidly with an agent, try slowing down the rate of requests.
I only asked 1 simple question about code to check model speed and there are endless errors where I get 500 erros and generating 5min responses to rename function in single file.
Hey All,
Thank you for the report. This has been escalated to our engineering team for investigation. To help us prioritize, please provide the details requested here.
To provide our developers with the necessary diagnostic data, we encourage you to use the in-app feedback tool (found via the Feedback icon in the top-right corner > Report Issue).
No specific error(s). Regardless of the model, the models stayed in a continuous loop, so I am forced to cancel the requests. This is problematic as in many cases they are able to partially complete the task, but because it needs to be cancelled. Understanding where they left it..is a nightmare.
This!! It says “Working, Generating, Loading“ most of the day, there are rare occasions when the models start working normally.
I’ve said this in my thread, but I’ll repeat.
GOOGLE, you’re giving out too many free generations to free/trial users, too many promotions like the students’ one. You’re constantly overloaded, especially when you release new models.
This is totally unfair for the paid customers. I bought an Ultra subscription for Antigravity, but since I bought it, the tool is unusable, and I just had to buy your rival’s (Claude Code) subscription and continue working on my projects with it. Shame on you, Google.
So many problems lately, AI studio down for 2 weeks, Antigravity unstable for the same amount of time, for what? Giving free generations to people who don’t really create anything and mostly play around?
Paid users have ongoing projects. We either pay you or we pay someone else, and you can continue being a buggy playground.
no matter on which plan are you its there side problem, model are not responding, taking lot of time, almost infinite, sometime its failed to send prompts